Built site for gh-pages

This commit is contained in:
Quarto GHA Workflow Runner
2025-02-25 09:11:03 +00:00
parent 4c1ebd7402
commit efdea0e9ff
48 changed files with 10359 additions and 5198 deletions

View File

@@ -8,7 +8,7 @@
<meta name="description" content="A comprehensive guide for using Axolotl on distributed systems with AMD GPUs">
<title>Training with AMD GPUs on HPC Systems Axolotl</title>
<title>AMD GPUs on HPC Systems Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/amd_hpc.html">Training with AMD GPUs on HPC Systems</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/amd_hpc.html">AMD GPUs on HPC Systems</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -350,9 +453,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/amd_hpc.html">Training with AMD GPUs on HPC Systems</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/amd_hpc.html">AMD GPUs on HPC Systems</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Training with AMD GPUs on HPC Systems</h1>
<h1 class="title">AMD GPUs on HPC Systems</h1>
</div>
<div>

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">Batch size vs Gradient accumulation</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Batch size vs Gradient accumulation</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +207,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -296,7 +399,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Batch size vs Gradient accumulation</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Batch size vs Gradient accumulation</h1>
</div>

View File

@@ -7,7 +7,7 @@
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<title>cli Axolotl</title>
<title>CLI Reference Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/cli.html">CLI Reference</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,82 +240,252 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
<div id="quarto-sidebar-glass" class="quarto-sidebar-collapse-item" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item"></div>
<!-- margin-sidebar -->
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
<nav id="TOC" role="doc-toc" class="toc-active">
<nav id="TOC" role="doc-toc" class="toc-active" data-toc-expanded="2">
<h2 id="toc-title">On this page</h2>
<ul>
<li><a href="#axolotl-cli-documentation" id="toc-axolotl-cli-documentation" class="nav-link active" data-scroll-target="#axolotl-cli-documentation">Axolotl CLI Documentation</a>
<li><a href="#basic-commands" id="toc-basic-commands" class="nav-link active" data-scroll-target="#basic-commands">Basic Commands</a></li>
<li><a href="#command-reference" id="toc-command-reference" class="nav-link" data-scroll-target="#command-reference">Command Reference</a>
<ul class="collapse">
<li><a href="#table-of-contents" id="toc-table-of-contents" class="nav-link" data-scroll-target="#table-of-contents">Table of Contents</a></li>
<li><a href="#basic-commands" id="toc-basic-commands" class="nav-link" data-scroll-target="#basic-commands">Basic Commands</a></li>
<li><a href="#command-reference" id="toc-command-reference" class="nav-link" data-scroll-target="#command-reference">Command Reference</a></li>
<li><a href="#fetch" id="toc-fetch" class="nav-link" data-scroll-target="#fetch">fetch</a></li>
<li><a href="#preprocess" id="toc-preprocess" class="nav-link" data-scroll-target="#preprocess">preprocess</a></li>
<li><a href="#train" id="toc-train" class="nav-link" data-scroll-target="#train">train</a></li>
<li><a href="#inference" id="toc-inference" class="nav-link" data-scroll-target="#inference">inference</a></li>
<li><a href="#merge-lora" id="toc-merge-lora" class="nav-link" data-scroll-target="#merge-lora">merge-lora</a></li>
<li><a href="#merge-sharded-fsdp-weights" id="toc-merge-sharded-fsdp-weights" class="nav-link" data-scroll-target="#merge-sharded-fsdp-weights">merge-sharded-fsdp-weights</a></li>
<li><a href="#evaluate" id="toc-evaluate" class="nav-link" data-scroll-target="#evaluate">evaluate</a></li>
<li><a href="#lm-eval" id="toc-lm-eval" class="nav-link" data-scroll-target="#lm-eval">lm-eval</a></li>
</ul></li>
<li><a href="#legacy-cli-usage" id="toc-legacy-cli-usage" class="nav-link" data-scroll-target="#legacy-cli-usage">Legacy CLI Usage</a></li>
<li><a href="#remote-compute-with-modal-cloud" id="toc-remote-compute-with-modal-cloud" class="nav-link" data-scroll-target="#remote-compute-with-modal-cloud">Remote Compute with Modal Cloud</a></li>
<li><a href="#remote-compute-with-modal-cloud" id="toc-remote-compute-with-modal-cloud" class="nav-link" data-scroll-target="#remote-compute-with-modal-cloud">Remote Compute with Modal Cloud</a>
<ul class="collapse">
<li><a href="#cloud-configuration" id="toc-cloud-configuration" class="nav-link" data-scroll-target="#cloud-configuration">Cloud Configuration</a></li>
<li><a href="#running-on-modal-cloud" id="toc-running-on-modal-cloud" class="nav-link" data-scroll-target="#running-on-modal-cloud">Running on Modal Cloud</a></li>
<li><a href="#cloud-configuration-options" id="toc-cloud-configuration-options" class="nav-link" data-scroll-target="#cloud-configuration-options">Cloud Configuration Options</a></li>
</ul></li>
</ul>
</nav>
</div>
<!-- main -->
<main class="content" id="quarto-document-content"><header id="title-block-header" class="quarto-title-block"></header>
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/cli.html">CLI Reference</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">CLI Reference</h1>
</div>
<div class="quarto-title-meta">
</div>
</header>
<section id="axolotl-cli-documentation" class="level1">
<h1>Axolotl CLI Documentation</h1>
<p>The Axolotl CLI provides a streamlined interface for training and fine-tuning large language models. This guide covers the CLI commands, their usage, and common examples.</p>
<section id="table-of-contents" class="level3">
<h3 class="anchored" data-anchor-id="table-of-contents">Table of Contents</h3>
<ul>
<li>Basic Commands</li>
<li>Command Reference
<ul>
<li>fetch</li>
<li>preprocess</li>
<li>train</li>
<li>inference</li>
<li>merge-lora</li>
<li>merge-sharded-fsdp-weights</li>
<li>evaluate</li>
<li>lm-eval</li>
</ul></li>
<li>Legacy CLI Usage</li>
<li>Remote Compute with Modal Cloud
<ul>
<li>Cloud Configuration</li>
<li>Running on Modal Cloud</li>
<li>Cloud Configuration Options</li>
</ul></li>
</ul>
</section>
<section id="basic-commands" class="level3">
<h3 class="anchored" data-anchor-id="basic-commands">Basic Commands</h3>
<section id="basic-commands" class="level2">
<h2 class="anchored" data-anchor-id="basic-commands">Basic Commands</h2>
<p>All Axolotl commands follow this general structure:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> <span class="op">&lt;</span>command<span class="op">&gt;</span> <span class="pp">[</span><span class="ss">config.yml</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">options</span><span class="pp">]</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>The config file can be local or a URL to a raw YAML file.</p>
</section>
<section id="command-reference" class="level3">
<h3 class="anchored" data-anchor-id="command-reference">Command Reference</h3>
<section id="fetch" class="level4">
<h4 class="anchored" data-anchor-id="fetch">fetch</h4>
<section id="command-reference" class="level2">
<h2 class="anchored" data-anchor-id="command-reference">Command Reference</h2>
<section id="fetch" class="level3">
<h3 class="anchored" data-anchor-id="fetch">fetch</h3>
<p>Downloads example configurations and deepspeed configs to your local machine.</p>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Get example YAML files</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> fetch examples</span>
@@ -392,8 +496,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<span id="cb2-7"><a href="#cb2-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Specify custom destination</span></span>
<span id="cb2-8"><a href="#cb2-8" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> fetch examples <span class="at">--dest</span> path/to/folder</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="preprocess" class="level4">
<h4 class="anchored" data-anchor-id="preprocess">preprocess</h4>
<section id="preprocess" class="level3">
<h3 class="anchored" data-anchor-id="preprocess">preprocess</h3>
<p>Preprocesses and tokenizes your dataset before training. This is recommended for large datasets.</p>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic preprocessing</span></span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> preprocess config.yml</span>
@@ -410,8 +514,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<div class="sourceCode" id="cb4"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="fu">dataset_prepared_path</span><span class="kw">:</span><span class="at"> Local folder for saving preprocessed data</span></span>
<span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a><span class="fu">push_dataset_to_hub</span><span class="kw">:</span><span class="at"> HuggingFace repo to push preprocessed data (optional)</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="train" class="level4">
<h4 class="anchored" data-anchor-id="train">train</h4>
<section id="train" class="level3">
<h3 class="anchored" data-anchor-id="train">train</h3>
<p>Trains or fine-tunes a model using the configuration specified in your YAML file.</p>
<div class="sourceCode" id="cb5"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic training</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml</span>
@@ -427,134 +531,173 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<span id="cb5-12"><a href="#cb5-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-13"><a href="#cb5-13" aria-hidden="true" tabindex="-1"></a><span class="co"># Resume training from checkpoint</span></span>
<span id="cb5-14"><a href="#cb5-14" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--resume-from-checkpoint</span> path/to/checkpoint</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>It is possible to run sweeps over multiple hyperparameters by passing in a sweeps config.</p>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic training with sweeps</span></span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--sweep</span> path/to/sweep.yaml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Example sweep config:</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="fu">_</span><span class="kw">:</span></span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="co"> # This section is for dependent variables we need to fix</span></span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">load_in_8bit</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb7-4"><a href="#cb7-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">load_in_4bit</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb7-5"><a href="#cb7-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">adapter</span><span class="kw">:</span><span class="at"> lora</span></span>
<span id="cb7-6"><a href="#cb7-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">load_in_8bit</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb7-7"><a href="#cb7-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">load_in_4bit</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb7-8"><a href="#cb7-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">adapter</span><span class="kw">:</span><span class="at"> lora</span></span>
<span id="cb7-9"><a href="#cb7-9" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-10"><a href="#cb7-10" aria-hidden="true" tabindex="-1"></a><span class="co"># These are independent variables</span></span>
<span id="cb7-11"><a href="#cb7-11" aria-hidden="true" tabindex="-1"></a><span class="fu">learning_rate</span><span class="kw">:</span><span class="at"> </span><span class="kw">[</span><span class="fl">0.0003</span><span class="kw">,</span><span class="at"> </span><span class="fl">0.0006</span><span class="kw">]</span></span>
<span id="cb7-12"><a href="#cb7-12" aria-hidden="true" tabindex="-1"></a><span class="fu">lora_r</span><span class="kw">:</span></span>
<span id="cb7-13"><a href="#cb7-13" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="dv">16</span></span>
<span id="cb7-14"><a href="#cb7-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="dv">32</span></span>
<span id="cb7-15"><a href="#cb7-15" aria-hidden="true" tabindex="-1"></a><span class="fu">lora_alpha</span><span class="kw">:</span></span>
<span id="cb7-16"><a href="#cb7-16" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="dv">16</span></span>
<span id="cb7-17"><a href="#cb7-17" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="dv">32</span></span>
<span id="cb7-18"><a href="#cb7-18" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="dv">64</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="inference" class="level4">
<h4 class="anchored" data-anchor-id="inference">inference</h4>
<section id="inference" class="level3">
<h3 class="anchored" data-anchor-id="inference">inference</h3>
<p>Runs inference using your trained model in either CLI or Gradio interface mode.</p>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="co"># CLI inference with LoRA</span></span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a><span class="co"># CLI inference with full model</span></span>
<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--base-model</span><span class="op">=</span><span class="st">"./completed-model"</span></span>
<span id="cb6-6"><a href="#cb6-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-7"><a href="#cb6-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Gradio web interface</span></span>
<span id="cb6-8"><a href="#cb6-8" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--gradio</span> <span class="dt">\</span></span>
<span id="cb6-9"><a href="#cb6-9" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb6-10"><a href="#cb6-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-11"><a href="#cb6-11" aria-hidden="true" tabindex="-1"></a><span class="co"># Inference with input from file</span></span>
<span id="cb6-12"><a href="#cb6-12" aria-hidden="true" tabindex="-1"></a><span class="fu">cat</span> prompt.txt <span class="kw">|</span> <span class="ex">axolotl</span> inference config.yml <span class="dt">\</span></span>
<span id="cb6-13"><a href="#cb6-13" aria-hidden="true" tabindex="-1"></a> <span class="at">--base-model</span><span class="op">=</span><span class="st">"./completed-model"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb8"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="co"># CLI inference with LoRA</span></span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb8-4"><a href="#cb8-4" aria-hidden="true" tabindex="-1"></a><span class="co"># CLI inference with full model</span></span>
<span id="cb8-5"><a href="#cb8-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--base-model</span><span class="op">=</span><span class="st">"./completed-model"</span></span>
<span id="cb8-6"><a href="#cb8-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb8-7"><a href="#cb8-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Gradio web interface</span></span>
<span id="cb8-8"><a href="#cb8-8" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference config.yml <span class="at">--gradio</span> <span class="dt">\</span></span>
<span id="cb8-9"><a href="#cb8-9" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb8-10"><a href="#cb8-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb8-11"><a href="#cb8-11" aria-hidden="true" tabindex="-1"></a><span class="co"># Inference with input from file</span></span>
<span id="cb8-12"><a href="#cb8-12" aria-hidden="true" tabindex="-1"></a><span class="fu">cat</span> prompt.txt <span class="kw">|</span> <span class="ex">axolotl</span> inference config.yml <span class="dt">\</span></span>
<span id="cb8-13"><a href="#cb8-13" aria-hidden="true" tabindex="-1"></a> <span class="at">--base-model</span><span class="op">=</span><span class="st">"./completed-model"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="merge-lora" class="level4">
<h4 class="anchored" data-anchor-id="merge-lora">merge-lora</h4>
<section id="merge-lora" class="level3">
<h3 class="anchored" data-anchor-id="merge-lora">merge-lora</h3>
<p>Merges trained LoRA adapters into the base model.</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic merge</span></span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-lora config.yml</span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-4"><a href="#cb7-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Specify LoRA directory (usually used with checkpoints)</span></span>
<span id="cb7-5"><a href="#cb7-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-lora config.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./lora-output/checkpoint-100"</span></span>
<span id="cb7-6"><a href="#cb7-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-7"><a href="#cb7-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Merge using CPU (if out of GPU memory)</span></span>
<span id="cb7-8"><a href="#cb7-8" aria-hidden="true" tabindex="-1"></a><span class="va">CUDA_VISIBLE_DEVICES</span><span class="op">=</span><span class="st">""</span> <span class="ex">axolotl</span> merge-lora config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Configuration options:</p>
<div class="sourceCode" id="cb8"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_memory_limit</span><span class="kw">:</span><span class="at"> Limit GPU memory usage</span></span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="fu">lora_on_cpu</span><span class="kw">:</span><span class="at"> Load LoRA weights on CPU</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="merge-sharded-fsdp-weights" class="level4">
<h4 class="anchored" data-anchor-id="merge-sharded-fsdp-weights">merge-sharded-fsdp-weights</h4>
<p>Merges sharded FSDP model checkpoints into a single combined checkpoint.</p>
<div class="sourceCode" id="cb9"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic merge</span></span>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-sharded-fsdp-weights config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="evaluate" class="level4">
<h4 class="anchored" data-anchor-id="evaluate">evaluate</h4>
<p>Evaluates a models performance using metrics specified in the config.</p>
<div class="sourceCode" id="cb10"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic evaluation</span></span>
<span id="cb10-2"><a href="#cb10-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> evaluate config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="lm-eval" class="level4">
<h4 class="anchored" data-anchor-id="lm-eval">lm-eval</h4>
<p>Runs LM Evaluation Harness on your model.</p>
<div class="sourceCode" id="cb11"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb11-1"><a href="#cb11-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic evaluation</span></span>
<span id="cb11-2"><a href="#cb11-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> lm-eval config.yml</span>
<span id="cb11-3"><a href="#cb11-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-4"><a href="#cb11-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Evaluate specific tasks</span></span>
<span id="cb11-5"><a href="#cb11-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> lm-eval config.yml <span class="at">--tasks</span> arc_challenge,hellaswag</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-lora config.yml</span>
<span id="cb9-3"><a href="#cb9-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-4"><a href="#cb9-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Specify LoRA directory (usually used with checkpoints)</span></span>
<span id="cb9-5"><a href="#cb9-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-lora config.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./lora-output/checkpoint-100"</span></span>
<span id="cb9-6"><a href="#cb9-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-7"><a href="#cb9-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Merge using CPU (if out of GPU memory)</span></span>
<span id="cb9-8"><a href="#cb9-8" aria-hidden="true" tabindex="-1"></a><span class="va">CUDA_VISIBLE_DEVICES</span><span class="op">=</span><span class="st">""</span> <span class="ex">axolotl</span> merge-lora config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Configuration options:</p>
<div class="sourceCode" id="cb12"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb12-1"><a href="#cb12-1" aria-hidden="true" tabindex="-1"></a><span class="fu">lm_eval_tasks</span><span class="kw">:</span><span class="at"> List of tasks to evaluate</span></span>
<span id="cb12-2"><a href="#cb12-2" aria-hidden="true" tabindex="-1"></a><span class="fu">lm_eval_batch_size</span><span class="kw">:</span><span class="at"> Batch size for evaluation</span></span>
<span id="cb12-3"><a href="#cb12-3" aria-hidden="true" tabindex="-1"></a><span class="fu">output_dir</span><span class="kw">:</span><span class="at"> Directory to save evaluation results</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb10"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_memory_limit</span><span class="kw">:</span><span class="at"> Limit GPU memory usage</span></span>
<span id="cb10-2"><a href="#cb10-2" aria-hidden="true" tabindex="-1"></a><span class="fu">lora_on_cpu</span><span class="kw">:</span><span class="at"> Load LoRA weights on CPU</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="merge-sharded-fsdp-weights" class="level3">
<h3 class="anchored" data-anchor-id="merge-sharded-fsdp-weights">merge-sharded-fsdp-weights</h3>
<p>Merges sharded FSDP model checkpoints into a single combined checkpoint.</p>
<div class="sourceCode" id="cb11"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb11-1"><a href="#cb11-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic merge</span></span>
<span id="cb11-2"><a href="#cb11-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> merge-sharded-fsdp-weights config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="evaluate" class="level3">
<h3 class="anchored" data-anchor-id="evaluate">evaluate</h3>
<p>Evaluates a models performance using metrics specified in the config.</p>
<div class="sourceCode" id="cb12"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb12-1"><a href="#cb12-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic evaluation</span></span>
<span id="cb12-2"><a href="#cb12-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> evaluate config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="lm-eval" class="level3">
<h3 class="anchored" data-anchor-id="lm-eval">lm-eval</h3>
<p>Runs LM Evaluation Harness on your model.</p>
<div class="sourceCode" id="cb13"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb13-1"><a href="#cb13-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Basic evaluation</span></span>
<span id="cb13-2"><a href="#cb13-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> lm-eval config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Configuration options:</p>
<div class="sourceCode" id="cb14"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb14-1"><a href="#cb14-1" aria-hidden="true" tabindex="-1"></a><span class="co"># List of tasks to evaluate</span></span>
<span id="cb14-2"><a href="#cb14-2" aria-hidden="true" tabindex="-1"></a><span class="fu">lm_eval_tasks</span><span class="kw">:</span></span>
<span id="cb14-3"><a href="#cb14-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> arc_challenge</span></span>
<span id="cb14-4"><a href="#cb14-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> hellaswag</span></span>
<span id="cb14-5"><a href="#cb14-5" aria-hidden="true" tabindex="-1"></a><span class="fu">lm_eval_batch_size</span><span class="kw">:</span><span class="co"> # Batch size for evaluation</span></span>
<span id="cb14-6"><a href="#cb14-6" aria-hidden="true" tabindex="-1"></a><span class="fu">output_dir</span><span class="kw">:</span><span class="co"> # Directory to save evaluation results</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
<section id="legacy-cli-usage" class="level3">
<h3 class="anchored" data-anchor-id="legacy-cli-usage">Legacy CLI Usage</h3>
<section id="legacy-cli-usage" class="level2">
<h2 class="anchored" data-anchor-id="legacy-cli-usage">Legacy CLI Usage</h2>
<p>While the new Click-based CLI is preferred, Axolotl still supports the legacy module-based CLI:</p>
<div class="sourceCode" id="cb13"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb13-1"><a href="#cb13-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Preprocess</span></span>
<span id="cb13-2"><a href="#cb13-2" aria-hidden="true" tabindex="-1"></a><span class="ex">python</span> <span class="at">-m</span> axolotl.cli.preprocess config.yml</span>
<span id="cb13-3"><a href="#cb13-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb13-4"><a href="#cb13-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Train</span></span>
<span id="cb13-5"><a href="#cb13-5" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.train config.yml</span>
<span id="cb13-6"><a href="#cb13-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb13-7"><a href="#cb13-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Inference</span></span>
<span id="cb13-8"><a href="#cb13-8" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.inference config.yml <span class="dt">\</span></span>
<span id="cb13-9"><a href="#cb13-9" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora_model_dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb13-10"><a href="#cb13-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb13-11"><a href="#cb13-11" aria-hidden="true" tabindex="-1"></a><span class="co"># Gradio interface</span></span>
<span id="cb13-12"><a href="#cb13-12" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.inference config.yml <span class="dt">\</span></span>
<span id="cb13-13"><a href="#cb13-13" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora_model_dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span> <span class="at">--gradio</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="remote-compute-with-modal-cloud" class="level3">
<h3 class="anchored" data-anchor-id="remote-compute-with-modal-cloud">Remote Compute with Modal Cloud</h3>
<p>Axolotl supports running training and inference workloads on Modal cloud infrastructure. This is configured using a cloud YAML file alongside your regular Axolotl config.</p>
<section id="cloud-configuration" class="level4">
<h4 class="anchored" data-anchor-id="cloud-configuration">Cloud Configuration</h4>
<p>Create a cloud config YAML with your Modal settings:</p>
<div class="sourceCode" id="cb14"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb14-1"><a href="#cb14-1" aria-hidden="true" tabindex="-1"></a><span class="co"># cloud_config.yml</span></span>
<span id="cb14-2"><a href="#cb14-2" aria-hidden="true" tabindex="-1"></a><span class="fu">provider</span><span class="kw">:</span><span class="at"> modal</span></span>
<span id="cb14-3"><a href="#cb14-3" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu</span><span class="kw">:</span><span class="at"> a100</span><span class="co"> # Supported: l40s, a100-40gb, a100-80gb, a10g, h100, t4, l4</span></span>
<span id="cb14-4"><a href="#cb14-4" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_count</span><span class="kw">:</span><span class="at"> </span><span class="dv">1</span><span class="co"> # Number of GPUs to use</span></span>
<span id="cb14-5"><a href="#cb14-5" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout</span><span class="kw">:</span><span class="at"> </span><span class="dv">86400</span><span class="co"> # Maximum runtime in seconds (24 hours)</span></span>
<span id="cb14-6"><a href="#cb14-6" aria-hidden="true" tabindex="-1"></a><span class="fu">branch</span><span class="kw">:</span><span class="at"> main</span><span class="co"> # Git branch to use (optional)</span></span>
<span id="cb14-7"><a href="#cb14-7" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb14-8"><a href="#cb14-8" aria-hidden="true" tabindex="-1"></a><span class="fu">volumes</span><span class="kw">:</span><span class="co"> # Persistent storage volumes</span></span>
<span id="cb14-9"><a href="#cb14-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">name</span><span class="kw">:</span><span class="at"> axolotl-cache</span></span>
<span id="cb14-10"><a href="#cb14-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">mount</span><span class="kw">:</span><span class="at"> /workspace/cache</span></span>
<span id="cb14-11"><a href="#cb14-11" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb14-12"><a href="#cb14-12" aria-hidden="true" tabindex="-1"></a><span class="fu">env</span><span class="kw">:</span><span class="co"> # Environment variables</span></span>
<span id="cb14-13"><a href="#cb14-13" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> WANDB_API_KEY</span></span>
<span id="cb14-14"><a href="#cb14-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> HF_TOKEN</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="running-on-modal-cloud" class="level4">
<h4 class="anchored" data-anchor-id="running-on-modal-cloud">Running on Modal Cloud</h4>
<p>Commands that support the cloud flag:</p>
<div class="sourceCode" id="cb15"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb15-1"><a href="#cb15-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Preprocess on cloud</span></span>
<span id="cb15-2"><a href="#cb15-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> preprocess config.yml <span class="at">--cloud</span> cloud_config.yml</span>
<div class="sourceCode" id="cb15"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb15-1"><a href="#cb15-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Preprocess</span></span>
<span id="cb15-2"><a href="#cb15-2" aria-hidden="true" tabindex="-1"></a><span class="ex">python</span> <span class="at">-m</span> axolotl.cli.preprocess config.yml</span>
<span id="cb15-3"><a href="#cb15-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb15-4"><a href="#cb15-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Train on cloud</span></span>
<span id="cb15-5"><a href="#cb15-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--cloud</span> cloud_config.yml</span>
<span id="cb15-4"><a href="#cb15-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Train</span></span>
<span id="cb15-5"><a href="#cb15-5" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.train config.yml</span>
<span id="cb15-6"><a href="#cb15-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb15-7"><a href="#cb15-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Train without accelerate on cloud</span></span>
<span id="cb15-8"><a href="#cb15-8" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--cloud</span> cloud_config.yml <span class="at">--no-accelerate</span></span>
<span id="cb15-9"><a href="#cb15-9" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb15-10"><a href="#cb15-10" aria-hidden="true" tabindex="-1"></a><span class="co"># Run lm-eval on cloud</span></span>
<span id="cb15-11"><a href="#cb15-11" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> lm-eval config.yml <span class="at">--cloud</span> cloud_config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<span id="cb15-7"><a href="#cb15-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Inference</span></span>
<span id="cb15-8"><a href="#cb15-8" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.inference config.yml <span class="dt">\</span></span>
<span id="cb15-9"><a href="#cb15-9" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora_model_dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span>
<span id="cb15-10"><a href="#cb15-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb15-11"><a href="#cb15-11" aria-hidden="true" tabindex="-1"></a><span class="co"># Gradio interface</span></span>
<span id="cb15-12"><a href="#cb15-12" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.inference config.yml <span class="dt">\</span></span>
<span id="cb15-13"><a href="#cb15-13" aria-hidden="true" tabindex="-1"></a> <span class="at">--lora_model_dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span> <span class="at">--gradio</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="callout callout-style-default callout-important callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Important
</div>
</div>
<div class="callout-body-container callout-body">
<p>When overriding CLI parameters in the legacy CLI, use same notation as in yaml file (e.g., <code>--lora_model_dir</code>).</p>
<p><strong>Note:</strong> This differs from the new Click-based CLI, which uses dash notation (e.g., <code>--lora-model-dir</code>). Keep this in mind if youre referencing newer documentation or switching between CLI versions.</p>
</div>
</div>
</section>
<section id="cloud-configuration-options" class="level4">
<h4 class="anchored" data-anchor-id="cloud-configuration-options">Cloud Configuration Options</h4>
<div class="sourceCode" id="cb16"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb16-1"><a href="#cb16-1" aria-hidden="true" tabindex="-1"></a><span class="fu">provider</span><span class="kw">:</span><span class="at"> compute provider, currently only `modal` is supported</span></span>
<span id="cb16-2"><a href="#cb16-2" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu</span><span class="kw">:</span><span class="at"> GPU type to use</span></span>
<span id="cb16-3"><a href="#cb16-3" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_count</span><span class="kw">:</span><span class="at"> Number of GPUs (default: 1)</span></span>
<span id="cb16-4"><a href="#cb16-4" aria-hidden="true" tabindex="-1"></a><span class="fu">memory</span><span class="kw">:</span><span class="at"> RAM in GB (default: 128)</span></span>
<span id="cb16-5"><a href="#cb16-5" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout</span><span class="kw">:</span><span class="at"> Maximum runtime in seconds</span></span>
<span id="cb16-6"><a href="#cb16-6" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout_preprocess</span><span class="kw">:</span><span class="at"> Preprocessing timeout</span></span>
<span id="cb16-7"><a href="#cb16-7" aria-hidden="true" tabindex="-1"></a><span class="fu">branch</span><span class="kw">:</span><span class="at"> Git branch to use</span></span>
<span id="cb16-8"><a href="#cb16-8" aria-hidden="true" tabindex="-1"></a><span class="fu">docker_tag</span><span class="kw">:</span><span class="at"> Custom Docker image tag</span></span>
<span id="cb16-9"><a href="#cb16-9" aria-hidden="true" tabindex="-1"></a><span class="fu">volumes</span><span class="kw">:</span><span class="at"> List of persistent storage volumes</span></span>
<span id="cb16-10"><a href="#cb16-10" aria-hidden="true" tabindex="-1"></a><span class="fu">env</span><span class="kw">:</span><span class="at"> Environment variables to pass</span></span>
<span id="cb16-11"><a href="#cb16-11" aria-hidden="true" tabindex="-1"></a><span class="fu">secrets</span><span class="kw">:</span><span class="at"> Secrets to inject</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<section id="remote-compute-with-modal-cloud" class="level2">
<h2 class="anchored" data-anchor-id="remote-compute-with-modal-cloud">Remote Compute with Modal Cloud</h2>
<p>Axolotl supports running training and inference workloads on Modal cloud infrastructure. This is configured using a cloud YAML file alongside your regular Axolotl config.</p>
<section id="cloud-configuration" class="level3">
<h3 class="anchored" data-anchor-id="cloud-configuration">Cloud Configuration</h3>
<p>Create a cloud config YAML with your Modal settings:</p>
<div class="sourceCode" id="cb16"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb16-1"><a href="#cb16-1" aria-hidden="true" tabindex="-1"></a><span class="co"># cloud_config.yml</span></span>
<span id="cb16-2"><a href="#cb16-2" aria-hidden="true" tabindex="-1"></a><span class="fu">provider</span><span class="kw">:</span><span class="at"> modal</span></span>
<span id="cb16-3"><a href="#cb16-3" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu</span><span class="kw">:</span><span class="at"> a100</span><span class="co"> # Supported: l40s, a100-40gb, a100-80gb, a10g, h100, t4, l4</span></span>
<span id="cb16-4"><a href="#cb16-4" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_count</span><span class="kw">:</span><span class="at"> </span><span class="dv">1</span><span class="co"> # Number of GPUs to use</span></span>
<span id="cb16-5"><a href="#cb16-5" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout</span><span class="kw">:</span><span class="at"> </span><span class="dv">86400</span><span class="co"> # Maximum runtime in seconds (24 hours)</span></span>
<span id="cb16-6"><a href="#cb16-6" aria-hidden="true" tabindex="-1"></a><span class="fu">branch</span><span class="kw">:</span><span class="at"> main</span><span class="co"> # Git branch to use (optional)</span></span>
<span id="cb16-7"><a href="#cb16-7" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb16-8"><a href="#cb16-8" aria-hidden="true" tabindex="-1"></a><span class="fu">volumes</span><span class="kw">:</span><span class="co"> # Persistent storage volumes</span></span>
<span id="cb16-9"><a href="#cb16-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">name</span><span class="kw">:</span><span class="at"> axolotl-cache</span></span>
<span id="cb16-10"><a href="#cb16-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">mount</span><span class="kw">:</span><span class="at"> /workspace/cache</span></span>
<span id="cb16-11"><a href="#cb16-11" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">name</span><span class="kw">:</span><span class="at"> axolotl-data</span></span>
<span id="cb16-12"><a href="#cb16-12" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">mount</span><span class="kw">:</span><span class="at"> /workspace/data</span></span>
<span id="cb16-13"><a href="#cb16-13" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">name</span><span class="kw">:</span><span class="at"> axolotl-artifacts</span></span>
<span id="cb16-14"><a href="#cb16-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">mount</span><span class="kw">:</span><span class="at"> /workspace/artifacts</span></span>
<span id="cb16-15"><a href="#cb16-15" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb16-16"><a href="#cb16-16" aria-hidden="true" tabindex="-1"></a><span class="fu">env</span><span class="kw">:</span><span class="co"> # Environment variables</span></span>
<span id="cb16-17"><a href="#cb16-17" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> WANDB_API_KEY</span></span>
<span id="cb16-18"><a href="#cb16-18" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> HF_TOKEN</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="running-on-modal-cloud" class="level3">
<h3 class="anchored" data-anchor-id="running-on-modal-cloud">Running on Modal Cloud</h3>
<p>Commands that support the cloud flag:</p>
<div class="sourceCode" id="cb17"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb17-1"><a href="#cb17-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Preprocess on cloud</span></span>
<span id="cb17-2"><a href="#cb17-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> preprocess config.yml <span class="at">--cloud</span> cloud_config.yml</span>
<span id="cb17-3"><a href="#cb17-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb17-4"><a href="#cb17-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Train on cloud</span></span>
<span id="cb17-5"><a href="#cb17-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--cloud</span> cloud_config.yml</span>
<span id="cb17-6"><a href="#cb17-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb17-7"><a href="#cb17-7" aria-hidden="true" tabindex="-1"></a><span class="co"># Train without accelerate on cloud</span></span>
<span id="cb17-8"><a href="#cb17-8" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--cloud</span> cloud_config.yml <span class="at">--no-accelerate</span></span>
<span id="cb17-9"><a href="#cb17-9" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb17-10"><a href="#cb17-10" aria-hidden="true" tabindex="-1"></a><span class="co"># Run lm-eval on cloud</span></span>
<span id="cb17-11"><a href="#cb17-11" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> lm-eval config.yml <span class="at">--cloud</span> cloud_config.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="cloud-configuration-options" class="level3">
<h3 class="anchored" data-anchor-id="cloud-configuration-options">Cloud Configuration Options</h3>
<div class="sourceCode" id="cb18"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb18-1"><a href="#cb18-1" aria-hidden="true" tabindex="-1"></a><span class="fu">provider</span><span class="kw">:</span><span class="co"> # compute provider, currently only `modal` is supported</span></span>
<span id="cb18-2"><a href="#cb18-2" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu</span><span class="kw">:</span><span class="co"> # GPU type to use</span></span>
<span id="cb18-3"><a href="#cb18-3" aria-hidden="true" tabindex="-1"></a><span class="fu">gpu_count</span><span class="kw">:</span><span class="co"> # Number of GPUs (default: 1)</span></span>
<span id="cb18-4"><a href="#cb18-4" aria-hidden="true" tabindex="-1"></a><span class="fu">memory</span><span class="kw">:</span><span class="co"> # RAM in GB (default: 128)</span></span>
<span id="cb18-5"><a href="#cb18-5" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout</span><span class="kw">:</span><span class="co"> # Maximum runtime in seconds</span></span>
<span id="cb18-6"><a href="#cb18-6" aria-hidden="true" tabindex="-1"></a><span class="fu">timeout_preprocess</span><span class="kw">:</span><span class="co"> # Preprocessing timeout</span></span>
<span id="cb18-7"><a href="#cb18-7" aria-hidden="true" tabindex="-1"></a><span class="fu">branch</span><span class="kw">:</span><span class="co"> # Git branch to use</span></span>
<span id="cb18-8"><a href="#cb18-8" aria-hidden="true" tabindex="-1"></a><span class="fu">docker_tag</span><span class="kw">:</span><span class="co"> # Custom Docker image tag</span></span>
<span id="cb18-9"><a href="#cb18-9" aria-hidden="true" tabindex="-1"></a><span class="fu">volumes</span><span class="kw">:</span><span class="co"> # List of persistent storage volumes</span></span>
<span id="cb18-10"><a href="#cb18-10" aria-hidden="true" tabindex="-1"></a><span class="fu">env</span><span class="kw">:</span><span class="co"> # Environment variables to pass</span></span>
<span id="cb18-11"><a href="#cb18-11" aria-hidden="true" tabindex="-1"></a><span class="fu">secrets</span><span class="kw">:</span><span class="co"> # Secrets to inject</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
</section>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>

File diff suppressed because it is too large Load Diff

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -368,7 +471,19 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<section id="sharegpt" class="level2">
<h2 class="anchored" data-anchor-id="sharegpt">sharegpt</h2>
<p>IMPORTANT: ShareGPT is deprecated!. Please see <a href="#chat_template">chat_template</a> section below.</p>
<div class="callout callout-style-default callout-important callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Important
</div>
</div>
<div class="callout-body-container callout-body">
<p>ShareGPT is deprecated!. Please see <a href="#chat_template">chat_template</a> section below.</p>
</div>
</div>
</section>
<section id="pygmalion" class="level2">
<h2 class="anchored" data-anchor-id="pygmalion">pygmalion</h2>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -339,6 +442,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li><a href="#setting-max_steps" id="toc-setting-max_steps" class="nav-link" data-scroll-target="#setting-max_steps">Setting max_steps</a></li>
<li><a href="#group_by_length" id="toc-group_by_length" class="nav-link" data-scroll-target="#group_by_length">Group_by_length</a></li>
</ul></li>
<li><a href="#reference" id="toc-reference" class="nav-link" data-scroll-target="#reference">Reference</a></li>
</ul></li>
<li><a href="#supervised-fine-tuning-sft" id="toc-supervised-fine-tuning-sft" class="nav-link" data-scroll-target="#supervised-fine-tuning-sft">Supervised fine-tuning (SFT)</a>
<ul>
@@ -353,9 +457,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li><a href="#choosing-a-chat_template" id="toc-choosing-a-chat_template" class="nav-link" data-scroll-target="#choosing-a-chat_template">Choosing a <code>chat_template</code></a></li>
<li><a href="#setting-chat_template-dataset-keys" id="toc-setting-chat_template-dataset-keys" class="nav-link" data-scroll-target="#setting-chat_template-dataset-keys">Setting <code>chat_template</code> dataset keys</a></li>
<li><a href="#handling-masking" id="toc-handling-masking" class="nav-link" data-scroll-target="#handling-masking">Handling masking</a></li>
</ul></li>
<li><a href="#applying-chat_template" id="toc-applying-chat_template" class="nav-link" data-scroll-target="#applying-chat_template">Applying <code>chat_template</code></a></li>
</ul></li>
<li><a href="#reference-1" id="toc-reference-1" class="nav-link" data-scroll-target="#reference-1">Reference</a></li>
</ul></li>
<li><a href="#instruction-dataset" id="toc-instruction-dataset" class="nav-link" data-scroll-target="#instruction-dataset">Instruction Dataset</a>
<ul>
<li><a href="#custom-instruct-prompt-format" id="toc-custom-instruct-prompt-format" class="nav-link" data-scroll-target="#custom-instruct-prompt-format">Custom Instruct Prompt Format</a></li>
@@ -396,7 +501,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<p>As there are a lot of available options in Axolotl, this guide aims to provide an simplify the user experience to choosing the proper choice.</p>
<p>Axolotl supports 3 kinds of training methods: pre-training, supervised fine-tuning, and preference-based post-training (e.g.&nbsp;DPO, ORPO, PRMs). Each method has their own dataset format which are described below.</p>
<section id="pre-training" class="level2">
<h2 class="anchored" data-anchor-id="pre-training"><a href="../../docs/dataset-formats/pretraining.html">Pre-training</a></h2>
<h2 class="anchored" data-anchor-id="pre-training">Pre-training</h2>
<p>When aiming to train on large corpora of text datasets, pre-training is your go-to choice. Due to the size of these datasets, downloading the entire-datasets before beginning training would be prohibitively time-consuming. Axolotl supports <a href="https://huggingface.co/docs/datasets/en/stream">streaming</a> to only load batches into memory at a time.</p>
<p>A sample format for a pre-training dataset is as follows:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span><span class="dt">"text"</span><span class="fu">:</span> <span class="st">"first row"</span><span class="fu">}</span></span>
@@ -463,6 +568,10 @@ Important
<p>It is recommended to leave this off if downloading from Hugging Face hub as it would download the entire dataset which can be very large.</p>
</section>
</section>
<section id="reference" class="level3">
<h3 class="anchored" data-anchor-id="reference">Reference</h3>
<p>Please see docs <a href="../../docs/dataset-formats/pretraining.html">here</a>.</p>
</section>
</section>
<section id="supervised-fine-tuning-sft" class="level2">
<h2 class="anchored" data-anchor-id="supervised-fine-tuning-sft">Supervised fine-tuning (SFT)</h2>
@@ -491,7 +600,7 @@ Tip
</div>
</div>
<section id="pre-tokenized-dataset" class="level3">
<h3 class="anchored" data-anchor-id="pre-tokenized-dataset"><a href="../../docs/dataset-formats/tokenized.html">Pre-Tokenized Dataset</a></h3>
<h3 class="anchored" data-anchor-id="pre-tokenized-dataset">Pre-Tokenized Dataset</h3>
<p>We suggest this approach when you want to bring your own tokenized dataset.</p>
<p>Axolotl expects the dataset to have three keys: - <code>input_ids</code>: from tokenizing formatted prompt - <code>attention_mask</code>: for masking padding. If you dont add padding, it would be equal to <code>len(input_ids) * [1]</code> - <code>labels</code>: this is the same as <code>input_ids</code>, however, if you want to mask certain tokens, you would set those indices to <code>-100</code>.</p>
<div class="callout callout-style-default callout-tip callout-titled">
@@ -524,9 +633,10 @@ Note
<p><code>type:</code> is empty!</p>
</div>
</div>
<p>Reference: <a href="../../docs/dataset-formats/tokenized.html">Pre-Tokenized Dataset Documentation</a>.</p>
</section>
<section id="template-free-dataset" class="level3">
<h3 class="anchored" data-anchor-id="template-free-dataset"><a href="../../docs/dataset-formats/template_free.html">Template Free Dataset</a></h3>
<h3 class="anchored" data-anchor-id="template-free-dataset">Template Free Dataset</h3>
<p>We reccomend this approach when you want granular control over the prompt formatting, special tokens, and masking, whilst letting Axolotl handle the tokenization. This is very useful if your dataset has unique prompts that differ across samples and where one single general template wouldnt suffice.</p>
<p>In the example below, you could see that there is no proper structure. At the same time, its very flexible as there are no constraints on how your prompt can look.</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
@@ -553,9 +663,10 @@ Note
<div class="sourceCode" id="cb8"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> A.jsonl</span></span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> input_output</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Reference: <a href="../../docs/dataset-formats/template_free.html">Template Free Documentation</a>.</p>
</section>
<section id="conversation-dataset" class="level3">
<h3 class="anchored" data-anchor-id="conversation-dataset"><a href="../../docs/dataset-formats/conversation.html">Conversation Dataset</a></h3>
<h3 class="anchored" data-anchor-id="conversation-dataset">Conversation Dataset</h3>
<p><code>conversation</code> messages are a list of messages which usually contain a <code>role</code> and <code>content</code> key.</p>
<div class="callout callout-style-default callout-tip callout-titled">
<div class="callout-header d-flex align-content-center">
@@ -618,7 +729,7 @@ The answer is 8.&lt;|im_end|&gt;</code></pre>
<p>Axolotl supports both as well as allowing customization of any kind of key.</p>
</section>
<section id="chat-template-usage" class="level4">
<h4 class="anchored" data-anchor-id="chat-template-usage"><a href="../../docs/dataset-formats/conversation.html#chat_template">Chat Template Usage</a></h4>
<h4 class="anchored" data-anchor-id="chat-template-usage">Chat Template Usage</h4>
<p>To properly use this method, it is important to identify three things:</p>
<ol type="1">
<li><p>Which <code>chat_template</code> would you use?</p></li>
@@ -673,39 +784,53 @@ The answer is 8.&lt;|im_end|&gt;</code></pre>
<span id="cb18-8"><a href="#cb18-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">user</span><span class="kw">:</span></span>
<span id="cb18-9"><a href="#cb18-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> human</span></span>
<span id="cb18-10"><a href="#cb18-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">narrator</span><span class="kw">:</span><span class="at"> </span><span class="kw">[</span><span class="st">"narrator"</span><span class="kw">]</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="callout callout-style-default callout-tip callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Tip
</div>
</div>
<div class="callout-body-container callout-body">
<p>As chat_templates may use hardcoded EOS/EOT tokens that are different from the tokenizers EOS, it is highly recommended to set them. For example, <code>ChatML</code> uses <code>&lt;|im_end|&gt;</code> to end turns.</p>
<div class="sourceCode" id="cb19"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb19-1"><a href="#cb19-1" aria-hidden="true" tabindex="-1"></a><span class="fu">special_tokens</span><span class="kw">:</span></span>
<span id="cb19-2"><a href="#cb19-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">eos_token</span><span class="kw">:</span><span class="at"> &lt;|im_end|&gt;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</div>
</div>
</section>
</section>
<section id="applying-chat_template" class="level4">
<h4 class="anchored" data-anchor-id="applying-chat_template">Applying <code>chat_template</code></h4>
<p>Once all the above steps are completed, you could combine all these configs together to form a bespoke configuration for your custom dataset. The final step would be to correctly set the EOS token in your config:</p>
<div class="sourceCode" id="cb19"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb19-1"><a href="#cb19-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb19-2"><a href="#cb19-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> A.jsonl</span></span>
<span id="cb19-3"><a href="#cb19-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chat_template</span></span>
<span id="cb19-4"><a href="#cb19-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-5"><a href="#cb19-5" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 1</span></span>
<span id="cb19-6"><a href="#cb19-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">chat_template</span><span class="kw">:</span><span class="at"> chatml</span></span>
<span id="cb19-7"><a href="#cb19-7" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-8"><a href="#cb19-8" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 2</span></span>
<span id="cb19-9"><a href="#cb19-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_messages</span><span class="kw">:</span><span class="at"> messages</span></span>
<span id="cb19-10"><a href="#cb19-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">message_property_mappings</span><span class="kw">:</span></span>
<span id="cb19-11"><a href="#cb19-11" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">role</span><span class="kw">:</span><span class="at"> role</span></span>
<span id="cb19-12"><a href="#cb19-12" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">content</span><span class="kw">:</span><span class="at"> content</span></span>
<span id="cb19-13"><a href="#cb19-13" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-14"><a href="#cb19-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">roles</span><span class="kw">:</span></span>
<span id="cb19-15"><a href="#cb19-15" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">assistant</span><span class="kw">:</span></span>
<span id="cb19-16"><a href="#cb19-16" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> gpt</span></span>
<span id="cb19-17"><a href="#cb19-17" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> model</span></span>
<span id="cb19-18"><a href="#cb19-18" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> assistant</span></span>
<span id="cb19-19"><a href="#cb19-19" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">user</span><span class="kw">:</span></span>
<span id="cb19-20"><a href="#cb19-20" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> human</span></span>
<span id="cb19-21"><a href="#cb19-21" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> user</span></span>
<span id="cb19-22"><a href="#cb19-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-23"><a href="#cb19-23" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 3</span></span>
<span id="cb19-24"><a href="#cb19-24" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">roles_to_train</span><span class="kw">:</span><span class="at"> </span><span class="kw">[</span><span class="st">"assistant"</span><span class="kw">]</span></span>
<span id="cb19-25"><a href="#cb19-25" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">train_on_eos</span><span class="kw">:</span><span class="at"> </span><span class="st">"turn"</span></span>
<span id="cb19-26"><a href="#cb19-26" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-27"><a href="#cb19-27" aria-hidden="true" tabindex="-1"></a><span class="fu">special_tokens</span><span class="kw">:</span></span>
<span id="cb19-28"><a href="#cb19-28" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">eos_token</span><span class="kw">:</span><span class="at"> &lt;|im_end|&gt;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<section id="applying-chat_template" class="level5">
<h5 class="anchored" data-anchor-id="applying-chat_template">Applying <code>chat_template</code></h5>
<p>Once all the above steps are completed, you could combine all these configs together to form a bespoke configuration for your custom dataset.</p>
<div class="sourceCode" id="cb20"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb20-1"><a href="#cb20-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb20-2"><a href="#cb20-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> A.jsonl</span></span>
<span id="cb20-3"><a href="#cb20-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chat_template</span></span>
<span id="cb20-4"><a href="#cb20-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb20-5"><a href="#cb20-5" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 1</span></span>
<span id="cb20-6"><a href="#cb20-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">chat_template</span><span class="kw">:</span><span class="at"> chatml</span></span>
<span id="cb20-7"><a href="#cb20-7" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb20-8"><a href="#cb20-8" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 2</span></span>
<span id="cb20-9"><a href="#cb20-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_messages</span><span class="kw">:</span><span class="at"> messages</span></span>
<span id="cb20-10"><a href="#cb20-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">message_property_mappings</span><span class="kw">:</span></span>
<span id="cb20-11"><a href="#cb20-11" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">role</span><span class="kw">:</span><span class="at"> role</span></span>
<span id="cb20-12"><a href="#cb20-12" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">content</span><span class="kw">:</span><span class="at"> content</span></span>
<span id="cb20-13"><a href="#cb20-13" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb20-14"><a href="#cb20-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">roles</span><span class="kw">:</span></span>
<span id="cb20-15"><a href="#cb20-15" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">assistant</span><span class="kw">:</span></span>
<span id="cb20-16"><a href="#cb20-16" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> gpt</span></span>
<span id="cb20-17"><a href="#cb20-17" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> model</span></span>
<span id="cb20-18"><a href="#cb20-18" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> assistant</span></span>
<span id="cb20-19"><a href="#cb20-19" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">user</span><span class="kw">:</span></span>
<span id="cb20-20"><a href="#cb20-20" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> human</span></span>
<span id="cb20-21"><a href="#cb20-21" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> user</span></span>
<span id="cb20-22"><a href="#cb20-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb20-23"><a href="#cb20-23" aria-hidden="true" tabindex="-1"></a><span class="co"> # step 3</span></span>
<span id="cb20-24"><a href="#cb20-24" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">roles_to_train</span><span class="kw">:</span><span class="at"> </span><span class="kw">[</span><span class="st">"assistant"</span><span class="kw">]</span></span>
<span id="cb20-25"><a href="#cb20-25" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">train_on_eos</span><span class="kw">:</span><span class="at"> </span><span class="st">"turn"</span></span>
<span id="cb20-26"><a href="#cb20-26" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb20-27"><a href="#cb20-27" aria-hidden="true" tabindex="-1"></a><span class="fu">special_tokens</span><span class="kw">:</span></span>
<span id="cb20-28"><a href="#cb20-28" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">eos_token</span><span class="kw">:</span><span class="at"> &lt;|im_end|&gt;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>If this config were to be applied to the sample dataset above, the output would look as such (which can be retrieved via <code>axolotl preprocess config.yaml --debug</code>):</p>
<pre><code>&lt;|im_start|&gt;(-100, 128256) user(-100, 882)
(-100, 198) Hi(-100, 13347) &lt;|im_end|&gt;(-100, 128257)
@@ -717,13 +842,31 @@ The answer is 8.&lt;|im_end|&gt;</code></pre>
(-100, 198) The(791, 791) answer(4320, 4320) is(374, 374) (220, 220) 8(23, 23) .(13, 13) &lt;|im_end|&gt;(128257, 128257)
(-100, 198)</code></pre>
<p>The first number refers to the label, the second refers to the <code>token_id</code>. For example, <code>-100</code> labels appear on non-assistant portions, meaning that they are masked during. For assistant portions, the label is the same as the <code>token_id</code>.</p>
<div class="callout callout-style-default callout-note callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Note
</div>
</div>
<div class="callout-body-container callout-body">
<p>If during <code>preprocess</code>, there are a lot of warnings of <code>Could not find content __ boundary</code>, please check the FAQ section for <a href="../../docs/faq.html#chat-templates">chat_templates</a>.</p>
</div>
</div>
</section>
</section>
<section id="reference-1" class="level4">
<h4 class="anchored" data-anchor-id="reference-1">Reference</h4>
<p>Please see docs <a href="../../docs/dataset-formats/conversation.html">here</a>.</p>
</section>
</section>
<section id="instruction-dataset" class="level3">
<h3 class="anchored" data-anchor-id="instruction-dataset"><a href="../../docs/dataset-formats/inst_tune.html">Instruction Dataset</a></h3>
<h3 class="anchored" data-anchor-id="instruction-dataset">Instruction Dataset</h3>
<p>Instruction datasets are used to train instruction-following models and comprise a prompt, containing an instruction, and a single response. In contrast to chat datasets which may be multi-turn, instruct datasets are typically single-turn.</p>
<p>An example is of a common format called Alpaca:</p>
<div class="sourceCode" id="cb21"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb21-1"><a href="#cb21-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span><span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"output"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb22"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb22-1"><a href="#cb22-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span><span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"output"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Using those keys, a prompt can be built based on it.</p>
<pre><code>Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
@@ -736,38 +879,40 @@ The answer is 8.&lt;|im_end|&gt;</code></pre>
### Response:
{output}</code></pre>
<p>This can be configured as such:</p>
<div class="sourceCode" id="cb23"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb23-1"><a href="#cb23-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb23-2"><a href="#cb23-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> A.jsonl</span></span>
<span id="cb23-3"><a href="#cb23-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> alpaca</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb24"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb24-1"><a href="#cb24-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb24-2"><a href="#cb24-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> A.jsonl</span></span>
<span id="cb24-3"><a href="#cb24-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> alpaca</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Axolotl supports many kinds of instruction dataset. All of them can be found here (https://axolotl-ai-cloud.github.io/axolotl/docs/dataset-formats/inst_tune.html) with their respective type and sample row format.</p>
<p>Reference: <a href="../../docs/dataset-formats/inst_tune.html">Instruction Dataset Documentation</a>.</p>
<section id="custom-instruct-prompt-format" class="level4">
<h4 class="anchored" data-anchor-id="custom-instruct-prompt-format">Custom Instruct Prompt Format</h4>
<p>Due to the myriad possibilities of instruction formats, Axolotl allows customizing your own instruction format without having to dive into the code directly.</p>
<p>In the example below, a sample row is used to output in <code>mistral_v1</code> format.</p>
<div class="sourceCode" id="cb24"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb24-1"><a href="#cb24-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span><span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"output"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb25"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb25-1"><a href="#cb25-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb25-2"><a href="#cb25-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> repo</span></span>
<span id="cb25-3"><a href="#cb25-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span></span>
<span id="cb25-4"><a href="#cb25-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">system_prompt</span><span class="kw">:</span><span class="at"> </span><span class="st">""</span></span>
<span id="cb25-5"><a href="#cb25-5" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb25-6"><a href="#cb25-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_system</span><span class="kw">:</span></span>
<span id="cb25-7"><a href="#cb25-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_instruction</span><span class="kw">:</span><span class="at"> input</span></span>
<span id="cb25-8"><a href="#cb25-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_input</span><span class="kw">:</span></span>
<span id="cb25-9"><a href="#cb25-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_output</span><span class="kw">:</span><span class="at"> output</span></span>
<span id="cb25-10"><a href="#cb25-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb25-11"><a href="#cb25-11" aria-hidden="true" tabindex="-1"></a><span class="co"> # multi-line example with input</span></span>
<span id="cb25-12"><a href="#cb25-12" aria-hidden="true" tabindex="-1"></a><span class="fu"> format</span><span class="kw">: </span><span class="ch">|-</span></span>
<span id="cb25-13"><a href="#cb25-13" aria-hidden="true" tabindex="-1"></a> [INST] {instruction} {input} [/INST]</span>
<span id="cb25-14"><a href="#cb25-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb25-15"><a href="#cb25-15" aria-hidden="true" tabindex="-1"></a><span class="co"> # single-line example without input</span></span>
<span id="cb25-16"><a href="#cb25-16" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">no_input_format</span><span class="kw">:</span><span class="at"> </span><span class="st">"[INST] {instruction} [/INST]"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb25"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb25-1"><a href="#cb25-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span><span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="dt">"output"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb26"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb26-1"><a href="#cb26-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb26-2"><a href="#cb26-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> repo</span></span>
<span id="cb26-3"><a href="#cb26-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span></span>
<span id="cb26-4"><a href="#cb26-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">system_prompt</span><span class="kw">:</span><span class="at"> </span><span class="st">""</span></span>
<span id="cb26-5"><a href="#cb26-5" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb26-6"><a href="#cb26-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_system</span><span class="kw">:</span></span>
<span id="cb26-7"><a href="#cb26-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_instruction</span><span class="kw">:</span><span class="at"> input</span></span>
<span id="cb26-8"><a href="#cb26-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_input</span><span class="kw">:</span></span>
<span id="cb26-9"><a href="#cb26-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">field_output</span><span class="kw">:</span><span class="at"> output</span></span>
<span id="cb26-10"><a href="#cb26-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb26-11"><a href="#cb26-11" aria-hidden="true" tabindex="-1"></a><span class="co"> # multi-line example with input</span></span>
<span id="cb26-12"><a href="#cb26-12" aria-hidden="true" tabindex="-1"></a><span class="fu"> format</span><span class="kw">: </span><span class="ch">|-</span></span>
<span id="cb26-13"><a href="#cb26-13" aria-hidden="true" tabindex="-1"></a> [INST] {instruction} {input} [/INST]</span>
<span id="cb26-14"><a href="#cb26-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb26-15"><a href="#cb26-15" aria-hidden="true" tabindex="-1"></a><span class="co"> # single-line example without input</span></span>
<span id="cb26-16"><a href="#cb26-16" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">no_input_format</span><span class="kw">:</span><span class="at"> </span><span class="st">"[INST] {instruction} [/INST]"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>The config sets that the <code>field_instruction</code> is actually named <code>input</code>, and the <code>field_input</code> is empty as we dont have an <code>input</code> in this sample. Generally, <code>instruction</code> can be thought as the question to the model, and <code>input</code> as the additional information with <code>output</code> being the response. It is not necessary to have an <code>input</code> nor <code>system</code>. In the end, the most important part is to understand what format you want it to look like and how you can customize this to your use case.</p>
<p>Reference: <a href="../../docs/dataset-formats/inst_tune.html#how-to-add-custom-prompt-format">Custom Instruct Prompt Format Documentation</a>.</p>
</section>
</section>
</section>
<section id="reinforcement-learning-from-human-feedback-rlhf" class="level2">
<h2 class="anchored" data-anchor-id="reinforcement-learning-from-human-feedback-rlhf">Reinforcement Learning from Human Feedback (RLHF)</h2>
<p>As there are multiple RLHF methods with their own dataset requirements. Please see <a href="../../docs/rlhf.html">RLHF datasets</a> documentation for more detail.</p>
<p>As there are multiple RLHF methods with their own dataset requirements. Please see <a href="../../docs/rlhf.html">RLHF documentation</a> for more detail.</p>
</section>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -385,8 +488,7 @@ Streaming is recommended for large datasets
<span id="cb2-5"><a href="#cb2-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">text_column</span><span class="kw">:</span><span class="co"> # column in dataset with the data, usually `text`</span></span>
<span id="cb2-6"><a href="#cb2-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> pretrain</span></span>
<span id="cb2-7"><a href="#cb2-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">trust_remote_code</span><span class="kw">:</span></span>
<span id="cb2-8"><a href="#cb2-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">skip</span><span class="kw">:</span><span class="co"> # number of rows of data to skip over from the beginning</span></span>
<span id="cb2-9"><a href="#cb2-9" aria-hidden="true" tabindex="-1"></a><span class="co">...</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<span id="cb2-8"><a href="#cb2-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">skip</span><span class="kw">:</span><span class="co"> # number of rows of data to skip over from the beginning</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</div>
</div>
</div>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>

View File

@@ -21,6 +21,40 @@ ul.task-list li input[type="checkbox"] {
margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */
vertical-align: middle;
}
/* CSS for syntax highlighting */
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
div.sourceCode { margin: 1em 0; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { display: inline-block; text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
{ counter-reset: source-line 0; }
pre.numberSource code > span
{ position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
{ content: counter(source-line);
position: relative; left: -1em; text-align: right; vertical-align: baseline;
border: none; display: inline-block;
-webkit-touch-callout: none; -webkit-user-select: none;
-khtml-user-select: none; -moz-user-select: none;
-ms-user-select: none; user-select: none;
padding: 0 4px; width: 4em;
}
pre.numberSource { margin-left: 3em; padding-left: 4px; }
div.sourceCode
{ }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
</style>
@@ -36,10 +70,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +113,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -117,7 +151,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +160,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +233,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,25 +241,211 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
<div id="quarto-sidebar-glass" class="quarto-sidebar-collapse-item" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item"></div>
<!-- margin-sidebar -->
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar zindex-bottom">
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
<nav id="TOC" role="doc-toc" class="toc-active">
<h2 id="toc-title">On this page</h2>
<ul>
<li><a href="#sec-background" id="toc-sec-background" class="nav-link active" data-scroll-target="#sec-background">Background</a>
<ul class="collapse">
<li><a href="#masking-inputs" id="toc-masking-inputs" class="nav-link" data-scroll-target="#masking-inputs">Masking Inputs</a></li>
<li><a href="#sec-you-may-not-want-prompt-templates" id="toc-sec-you-may-not-want-prompt-templates" class="nav-link" data-scroll-target="#sec-you-may-not-want-prompt-templates">You may not want prompt templates</a></li>
<li><a href="#sec-the-inputoutput-format" id="toc-sec-the-inputoutput-format" class="nav-link" data-scroll-target="#sec-the-inputoutput-format">The <code>input_output</code> format</a></li>
</ul></li>
<li><a href="#sec-usage" id="toc-sec-usage" class="nav-link" data-scroll-target="#sec-usage">Usage</a>
<ul class="collapse">
<li><a href="#sec-1-prepare-data" id="toc-sec-1-prepare-data" class="nav-link" data-scroll-target="#sec-1-prepare-data">1. Prepare Data</a></li>
<li><a href="#sec-2-use-type-inputoutput" id="toc-sec-2-use-type-inputoutput" class="nav-link" data-scroll-target="#sec-2-use-type-inputoutput">2. Use <code>type: input_output</code></a></li>
<li><a href="#sec-3-check-the-prompts" id="toc-sec-3-check-the-prompts" class="nav-link" data-scroll-target="#sec-3-check-the-prompts">3. Check the prompts</a></li>
</ul></li>
</ul>
</nav>
</div>
<!-- main -->
<main class="content" id="quarto-document-content">
@@ -320,9 +474,234 @@ ul.task-list li input[type="checkbox"] {
</header>
<p>See <a href="../../docs/input_output.html">these docs</a>.</p>
<section id="sec-background" class="level2">
<h2 class="anchored" data-anchor-id="sec-background">Background</h2>
<section id="masking-inputs" class="level3">
<h3 class="anchored" data-anchor-id="masking-inputs">Masking Inputs</h3>
<p>One of the most popular features of <a href="https://github.com/axolotl-ai-cloud/axolotl">axolotl</a> is setting the following configuration value:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>If you declare a <a href="https://github.com/axolotl-ai-cloud/axolotl?tab=readme-ov-file#dataset">dataset formats</a> such as <code>alpaca</code> or <code>chatml</code>, axolotl knows what is an input (i.e.&nbsp;human) vs.&nbsp;an output (i.e.&nbsp;the assistant) and masks the input labels so that your model can focus on predicting the outputs only.</p>
</section>
<section id="sec-you-may-not-want-prompt-templates" class="level3">
<h3 class="anchored" data-anchor-id="sec-you-may-not-want-prompt-templates">You may not want prompt templates</h3>
<p>However, there are many situations where you dont want to use one of these formats or templates. This is because they can:</p>
<ul>
<li>Add unnecessary boilerplate to your prompts.</li>
<li>Create artifacts like special delimiters <code>&lt;|im_start|&gt;</code> that can quickly become footguns if you dont include them correctly at inference time.</li>
<li>Enforce a <em>chat</em> interface when you do not want one. Sometimes you just want to fine-tune a model to a very specific task and do NOT want multi-turn conversations, roles, etc.</li>
<li>Limit you to only certain roles that the template allows.</li>
</ul>
</section>
<section id="sec-the-inputoutput-format" class="level3">
<h3 class="anchored" data-anchor-id="sec-the-inputoutput-format">The <code>input_output</code> format</h3>
<p>You can construct your prompts without a template by using the <code>input_output</code> format, by setting <code>type: input_output</code> in your configuration file like this:</p>
<p><strong>config.yml</strong></p>
<div class="sourceCode" id="cb2"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span><span class="co"> # Mask segments of your data</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> output.jsonl</span></span>
<span id="cb2-4"><a href="#cb2-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> input_output</span><span class="co"> # use template free prompt construction</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Unlike <code>type: completion</code>, which is also template-free, <code>type: input_output</code> allows you to mask segments of your text. More details on how this works are described below.</p>
</section>
</section>
<section id="sec-usage" class="level2">
<h2 class="anchored" data-anchor-id="sec-usage">Usage</h2>
<p>This is how you can use the <code>input_output</code> format:</p>
<section id="sec-1-prepare-data" class="level3">
<h3 class="anchored" data-anchor-id="sec-1-prepare-data">1. Prepare Data</h3>
<p>To use the <code>input_output</code> format, collect your data in the following format into a jsonl file (below is the first row from the file <code>output</code>.jsonl` pretty printed):</p>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="ex">$</span> head <span class="at">-n1</span> output.jsonl <span class="kw">|</span> <span class="ex">python</span> <span class="at">-m</span> json.tool</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="cell-output cell-output-stdout">
<pre><code>{
"segments": [
{
"label": true,
"text": "&lt;s&gt;Hello\n"
},
{
"label": true,
"text": "hi there!. "
},
{
"label": false,
"text": "goodbye "
},
{
"label": true,
"text": "farewell&lt;/s&gt;"
}
]
}</code></pre>
</div>
<p>Set <code>label:false</code> when you want to mask a segment of text so that the model isnt trained on it. Some things to keep in mind:</p>
<blockquote class="blockquote">
<p>[!IMPORTANT] 1. <strong>EOS, BOS, spaces, newlines etc. are entirely up to you. Axolotl concatenates all the segments as-is.</strong> The tokenizer doesnt add anything additional. Notice how I added spaces, newlines, <code>&lt;s&gt;</code> (BOS), and <code>&lt;/s&gt;</code> (EOS) myself. 2. Make sure you check the materialized output to validate that the prompt is getting assembled how you like.</p>
</blockquote>
</section>
<section id="sec-2-use-type-inputoutput" class="level3">
<h3 class="anchored" data-anchor-id="sec-2-use-type-inputoutput">2. Use <code>type: input_output</code></h3>
<p>Lets materialize data with our <code>output.jsonl</code> file by setting <code>type: input_output</code> in our axolotl config:</p>
<div class="sourceCode" id="cb5"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="co"># training_config.yaml</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="fu">base_model</span><span class="kw">:</span><span class="at"> mistralai/Mistral-7B-v0.1</span></span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a><span class="fu">data_seed</span><span class="kw">:</span><span class="at"> </span><span class="dv">49</span></span>
<span id="cb5-4"><a href="#cb5-4" aria-hidden="true" tabindex="-1"></a><span class="fu">seed</span><span class="kw">:</span><span class="at"> </span><span class="dv">49</span></span>
<span id="cb5-5"><a href="#cb5-5" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-6"><a href="#cb5-6" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb5-7"><a href="#cb5-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> output.jsonl</span></span>
<span id="cb5-8"><a href="#cb5-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> input_output</span></span>
<span id="cb5-9"><a href="#cb5-9" aria-hidden="true" tabindex="-1"></a><span class="fu">val_set_size</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.1</span></span>
<span id="cb5-10"><a href="#cb5-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-11"><a href="#cb5-11" aria-hidden="true" tabindex="-1"></a><span class="fu">sequence_len</span><span class="kw">:</span><span class="at"> </span><span class="dv">896</span></span>
<span id="cb5-12"><a href="#cb5-12" aria-hidden="true" tabindex="-1"></a><span class="fu">sample_packing</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb5-13"><a href="#cb5-13" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-14"><a href="#cb5-14" aria-hidden="true" tabindex="-1"></a><span class="fu">micro_batch_size</span><span class="kw">:</span><span class="at"> </span><span class="dv">2</span></span>
<span id="cb5-15"><a href="#cb5-15" aria-hidden="true" tabindex="-1"></a><span class="fu">gradient_accumulation_steps</span><span class="kw">:</span><span class="at"> </span><span class="dv">3</span></span>
<span id="cb5-16"><a href="#cb5-16" aria-hidden="true" tabindex="-1"></a><span class="fu">eval_batch_size</span><span class="kw">:</span><span class="at"> </span><span class="dv">2</span></span>
<span id="cb5-17"><a href="#cb5-17" aria-hidden="true" tabindex="-1"></a><span class="fu">num_epochs</span><span class="kw">:</span><span class="at"> </span><span class="dv">1</span></span>
<span id="cb5-18"><a href="#cb5-18" aria-hidden="true" tabindex="-1"></a><span class="fu">learning_rate</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.0002</span></span>
<span id="cb5-19"><a href="#cb5-19" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-20"><a href="#cb5-20" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb5-21"><a href="#cb5-21" aria-hidden="true" tabindex="-1"></a><span class="fu">special_tokens</span><span class="kw">:</span></span>
<span id="cb5-22"><a href="#cb5-22" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">bos_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;s&gt;"</span></span>
<span id="cb5-23"><a href="#cb5-23" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">eos_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;/s&gt;"</span></span>
<span id="cb5-24"><a href="#cb5-24" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">unk_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;unk&gt;"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>You can use the following command to materialize your data. The <code>--debug</code> flag will print the tokens, along with the labels so you can verify that the correct items are being ignored:</p>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> preprocess training_config.yaml <span class="at">--debug</span></span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a><span class="ex">...</span></span>
<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a><span class="ex">[2024-03-05</span> 23:36:46,969] <span class="pp">[</span><span class="ss">INFO</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">axolotl.check_example_labels:35</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">PID:607731</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">RANK:0</span><span class="pp">]</span> <span class="op">&lt;</span>s<span class="op">&gt;(</span><span class="ex">1,</span> 1<span class="op">)</span> Hello<span class="er">(</span><span class="ex">22557,</span> 22557<span class="kw">)</span></span>
<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a><span class="kw">(</span><span class="ex">13,</span> 13<span class="kw">)</span> <span class="ex">hi</span><span class="er">(</span><span class="ex">12014,</span> 12014<span class="kw">)</span> <span class="ex">there</span><span class="er">(</span><span class="ex">736,</span> 736<span class="kw">)</span> <span class="ex">!</span><span class="er">(</span><span class="ex">28808,</span> 28808<span class="kw">)</span> <span class="bu">.</span><span class="er">(</span><span class="ex">28723,</span> 28723<span class="kw">)</span> <span class="kw">(</span><span class="ex">28705,</span> 28705<span class="kw">)</span> <span class="ex">good</span><span class="er">(</span><span class="ex">-100,</span> 1179<span class="kw">)</span> <span class="ex">bye</span><span class="er">(</span><span class="ex">-100,</span> 17664<span class="kw">)</span> <span class="kw">(</span><span class="ex">-100,</span> 28705<span class="kw">)</span> <span class="ex">fare</span><span class="er">(</span><span class="ex">19111,</span> 19111<span class="kw">)</span> <span class="ex">well</span><span class="er">(</span><span class="ex">5458,</span> 5458<span class="kw">)</span> <span class="op">&lt;</span>/s<span class="op">&gt;(</span><span class="ex">2,</span> 2<span class="op">)</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>The format is <code>decoded_token</code>(<code>label</code>, <code>token_id</code>), for example, <code>&lt;s&gt;(1, 1)</code> means that the token is <code>&lt;s&gt;</code>, the label is <code>1</code> and the token_id is <code>1</code>. When the label is <code>-100</code> then that token is ignored for training.</p>
</section>
<section id="sec-3-check-the-prompts" class="level3">
<h3 class="anchored" data-anchor-id="sec-3-check-the-prompts">3. Check the prompts</h3>
<p>Here is another way to check the materialized output:</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> transformers <span class="im">import</span> AutoTokenizer</span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> datasets <span class="im">import</span> load_from_disk</span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> yaml</span>
<span id="cb7-4"><a href="#cb7-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-5"><a href="#cb7-5" aria-hidden="true" tabindex="-1"></a>directory <span class="op">=</span> <span class="op">!</span>ls last_run_prepared<span class="op">/</span></span>
<span id="cb7-6"><a href="#cb7-6" aria-hidden="true" tabindex="-1"></a><span class="cf">with</span> <span class="bu">open</span>(<span class="st">'training_config.yaml'</span>, <span class="st">'r'</span>) <span class="im">as</span> f:</span>
<span id="cb7-7"><a href="#cb7-7" aria-hidden="true" tabindex="-1"></a> cfg <span class="op">=</span> yaml.safe_load(f)</span>
<span id="cb7-8"><a href="#cb7-8" aria-hidden="true" tabindex="-1"></a>model_id <span class="op">=</span> cfg[<span class="st">'base_model'</span>]</span>
<span id="cb7-9"><a href="#cb7-9" aria-hidden="true" tabindex="-1"></a>tok <span class="op">=</span> AutoTokenizer.from_pretrained(model_id)</span>
<span id="cb7-10"><a href="#cb7-10" aria-hidden="true" tabindex="-1"></a>ds <span class="op">=</span> load_from_disk(<span class="ss">f'last_run_prepared/</span><span class="sc">{</span>directory[<span class="dv">0</span>]<span class="sc">}</span><span class="ss">/'</span>)</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb8"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="op">&gt;&gt;&gt;</span> row <span class="op">=</span> ds[<span class="dv">0</span>]</span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="op">&gt;&gt;&gt;</span> <span class="bu">print</span>(tok.decode(row[<span class="st">'input_ids'</span>]))</span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a><span class="op">&lt;</span>s<span class="op">&gt;</span> Hello</span>
<span id="cb8-4"><a href="#cb8-4" aria-hidden="true" tabindex="-1"></a> hi there<span class="op">!</span>. goodbye farewell<span class="op">&lt;/</span>s<span class="op">&gt;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>We can check that the right tokens are ignored by comparing the labels to each token:</p>
<div class="sourceCode" id="cb9"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> pandas <span class="im">as</span> pd</span>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a>pd.DataFrame([{<span class="st">'token'</span>: tok.decode(i), <span class="st">'label'</span>: l, <span class="st">'id'</span>:i} <span class="cf">for</span> i,l <span class="kw">in</span></span>
<span id="cb9-3"><a href="#cb9-3" aria-hidden="true" tabindex="-1"></a> <span class="bu">zip</span>(row[<span class="st">'input_ids'</span>], row[<span class="st">'labels'</span>])])</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<table class="caption-top table">
<thead>
<tr class="header">
<th>token</th>
<th>label</th>
<th>id</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td>0</td>
<td>&lt;s&gt;</td>
<td>1</td>
</tr>
<tr class="even">
<td>1</td>
<td>Hello</td>
<td>22557</td>
</tr>
<tr class="odd">
<td>2</td>
<td>\n</td>
<td>13</td>
</tr>
<tr class="even">
<td>3</td>
<td>hi</td>
<td>12014</td>
</tr>
<tr class="odd">
<td>4</td>
<td>there</td>
<td>736</td>
</tr>
<tr class="even">
<td>5</td>
<td>!</td>
<td>28808</td>
</tr>
<tr class="odd">
<td>6</td>
<td>.</td>
<td>28723</td>
</tr>
<tr class="even">
<td>7</td>
<td></td>
<td>28705</td>
</tr>
<tr class="odd">
<td>8</td>
<td>good</td>
<td>-100</td>
</tr>
<tr class="even">
<td>9</td>
<td>bye</td>
<td>-100</td>
</tr>
<tr class="odd">
<td>10</td>
<td></td>
<td>-100</td>
</tr>
<tr class="even">
<td>11</td>
<td>fare</td>
<td>19111</td>
</tr>
<tr class="odd">
<td>12</td>
<td>well</td>
<td>5458</td>
</tr>
<tr class="even">
<td>13</td>
<td>&lt;/s&gt;</td>
<td>2</td>
</tr>
</tbody>
</table>
<p>If we look at the input data, the above table seems correct! (The jsonl version is repeated below for reference):</p>
<div class="sourceCode" id="cb10"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="ex">$</span> head <span class="at">-n1</span> output.jsonl <span class="kw">|</span> <span class="ex">python</span> <span class="at">-m</span> json.tool</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="cell-output cell-output-stdout">
<pre><code>{
"segments": [
{
"label": true,
"text": "&lt;s&gt;Hello\n"
},
{
"label": true,
"text": "hi there!. "
},
{
"label": false,
"text": "goodbye "
},
{
"label": true,
"text": "farewell&lt;/s&gt;"
}
]
}</code></pre>
</div>
</section>
</section>
</main> <!-- /main -->
<script id="quarto-html-after-body" type="application/javascript">

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../../index.html" class="navbar-brand navbar-brand-logo">
<img src="../../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">Dataset Preprocessing</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/dataset_preprocessing.html">Dataset Preprocessing</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +207,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -296,7 +399,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/dataset_preprocessing.html">Dataset Preprocessing</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Dataset Preprocessing</h1>
</div>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/debugging.html">Debugging</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/debugging.html">Debugging</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -351,7 +454,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/debugging.html">Debugging</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/debugging.html">Debugging</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Debugging</h1>
</div>
@@ -402,21 +505,24 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<p>[!Important] All of these tips are incorporated into the <a href="#configuration">example configuration</a> for debugging with VSCode below.</p>
</blockquote>
<ol type="1">
<li><strong>Make sure you are using the latest version of axolotl</strong>: This project changes often and bugs get fixed fast. Check your git branch and make sure you have pulled the latest changes from <code>main</code>.</li>
<li><strong>Eliminate concurrency</strong>: Restrict the number of processes to 1 for both training and data preprocessing:
<li><p><strong>Make sure you are using the latest version of axolotl</strong>: This project changes often and bugs get fixed fast. Check your git branch and make sure you have pulled the latest changes from <code>main</code>.</p></li>
<li><p><strong>Eliminate concurrency</strong>: Restrict the number of processes to 1 for both training and data preprocessing:</p>
<ul>
<li>Set <code>CUDA_VISIBLE_DEVICES</code> to a single GPU, ex: <code>export CUDA_VISIBLE_DEVICES=0</code>.</li>
<li>Set <code>dataset_processes: 1</code> in your axolotl config or run the training command with <code>--dataset_processes=1</code>.</li>
</ul></li>
<li><strong>Use a small dataset</strong>: Construct or use a small dataset from HF Hub. When using a small dataset, you will often have to make sure <code>sample_packing: False</code> and <code>eval_sample_packing: False</code> to avoid errors. If you are in a pinch and dont have time to construct a small dataset but want to use from the HF Hub, you can shard the data (this will still tokenize the entire dataset, but will only use a fraction of the data for training. For example, to shard the dataset into 20 pieces, add the following to your axolotl config): <code>yaml dataset: ... shards: 20</code></li>
<li><strong>Use a small model</strong>: A good example of a small model is <a href="https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0">TinyLlama/TinyLlama-1.1B-Chat-v1.0</a>.</li>
<li><strong>Minimize iteration time</strong>: Make sure the training loop finishes as fast as possible, with these settings.
<li><p><strong>Use a small dataset</strong>: Construct or use a small dataset from HF Hub. When using a small dataset, you will often have to make sure <code>sample_packing: False</code> and <code>eval_sample_packing: False</code> to avoid errors. If you are in a pinch and dont have time to construct a small dataset but want to use from the HF Hub, you can shard the data (this will still tokenize the entire dataset, but will only use a fraction of the data for training. For example, to shard the dataset into 20 pieces, add the following to your axolotl config):</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb1-2"><a href="#cb1-2" aria-hidden="true" tabindex="-1"></a><span class="at"> ...</span></span>
<span id="cb1-3"><a href="#cb1-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">shards</span><span class="kw">:</span><span class="at"> </span><span class="dv">20</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div></li>
<li><p><strong>Use a small model</strong>: A good example of a small model is <a href="https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0">TinyLlama/TinyLlama-1.1B-Chat-v1.0</a>.</p></li>
<li><p><strong>Minimize iteration time</strong>: Make sure the training loop finishes as fast as possible, with these settings.</p>
<ul>
<li><code>micro_batch_size: 1</code></li>
<li><code>max_steps: 1</code></li>
<li><code>val_set_size: 0</code></li>
</ul></li>
<li><strong>Clear Caches:</strong> Axolotl caches certain steps and so does the underlying HuggingFace trainer. You may want to clear some of these caches when debugging.
<li><p><strong>Clear Caches:</strong> Axolotl caches certain steps and so does the underlying HuggingFace trainer. You may want to clear some of these caches when debugging.</p>
<ul>
<li>Data preprocessing: When debugging data preprocessing, which includes prompt template formation, you may want to delete the directory set in <code>dataset_prepared_path:</code> in your axolotl config. If you didnt set this value, the default is <code>last_run_prepared</code>.</li>
<li>HF Hub: If you are debugging data preprocessing, you should clear the relevant HF cache <a href="https://huggingface.co/docs/datasets/cache">HuggingFace cache</a>, by deleting the appropriate <code>~/.cache/huggingface/datasets/...</code> folder(s).</li>
@@ -429,9 +535,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<section id="background" class="level3">
<h3 class="anchored" data-anchor-id="background">Background</h3>
<p>The below example shows how to configure VSCode to debug data preprocessing of the <code>chat_template</code> format. This is the format used when you have the following in your axolotl config:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb1-2"><a href="#cb1-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> &lt;path to your chat_template formatted dataset&gt;</span><span class="co"> # example on HF Hub: fozziethebeat/alpaca_messages_2k_test</span></span>
<span id="cb1-3"><a href="#cb1-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chat_template</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb2"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> &lt;path to your chat_template formatted dataset&gt;</span><span class="co"> # example on HF Hub: fozziethebeat/alpaca_messages_2k_test</span></span>
<span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chat_template</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<blockquote class="blockquote">
<p>[!Important] If you are already familiar with advanced VSCode debugging, you can skip the below explanation and look at the files <a href="../.vscode/launch.json">.vscode/launch.json</a> and <a href="../.vscode/tasks.json">.vscode/tasks.json</a> for an example configuration.</p>
</blockquote>
@@ -442,8 +548,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<section id="setup" class="level3">
<h3 class="anchored" data-anchor-id="setup">Setup</h3>
<p>Make sure you have an <a href="https://setuptools.pypa.io/en/latest/userguide/development_mode.html">editable install</a> of Axolotl, which ensures that changes you make to the code are reflected at runtime. Run the following commands from the root of this project:</p>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install packaging</span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install <span class="at">--no-build-isolation</span> <span class="at">-e</span> <span class="st">'.[flash-attn,deepspeed]'</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install packaging</span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install <span class="at">--no-build-isolation</span> <span class="at">-e</span> <span class="st">'.[flash-attn,deepspeed]'</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<section id="remote-hosts" class="level4">
<h4 class="anchored" data-anchor-id="remote-hosts">Remote Hosts</h4>
<p>If you developing on a remote host, you can easily use VSCode to debug remotely. To do so, you will need to follow this <a href="https://code.visualstudio.com/docs/remote/ssh">remote - SSH guide</a>. You can also see the video below on <a href="#video---attaching-to-docker-on-remote-host">Docker and Remote SSH debugging</a>.</p>
@@ -453,38 +559,38 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<h3 class="anchored" data-anchor-id="configuration">Configuration</h3>
<p>The easiest way to get started is to modify the <a href="../.vscode/launch.json">.vscode/launch.json</a> file in this project. This is just an example configuration, so you may need to modify or copy it to suit your needs.</p>
<p>For example, to mimic the command <code>cd devtools &amp;&amp; CUDA_VISIBLE_DEVICES=0 accelerate launch -m axolotl.cli.train dev_chat_template.yml</code>, you would use the below configuration<a href="#fn1" class="footnote-ref" id="fnref1" role="doc-noteref"><sup>1</sup></a>. Note that we add additional flags that override the axolotl config and incorporate the tips above (see the comments). We also set the working directory to <code>devtools</code> and set the <code>env</code> variable <code>HF_HOME</code> to a temporary folder that is later partially deleted. This is because we want to delete the HF dataset cache before each run in order to ensure that the data preprocessing code is run from scratch.</p>
<pre class="jsonc"><code>// .vscode/launch.json
{
"version": "0.2.0",
"configurations": [
{
"name": "Debug axolotl prompt - chat_template",
"type": "python",
"module": "accelerate.commands.launch",
"request": "launch",
"args": [
"-m", "axolotl.cli.train", "dev_chat_template.yml",
// The flags below simplify debugging by overriding the axolotl config
// with the debugging tips above. Modify as needed.
"--dataset_processes=1", // limits data preprocessing to one process
"--max_steps=1", // limits training to just one step
"--batch_size=1", // minimizes batch size
"--micro_batch_size=1", // minimizes batch size
"--val_set_size=0", // disables validation
"--sample_packing=False", // disables sample packing which is necessary for small datasets
"--eval_sample_packing=False",// disables sample packing on eval set
"--dataset_prepared_path=temp_debug/axolotl_outputs/data", // send data outputs to a temp folder
"--output_dir=temp_debug/axolotl_outputs/model" // send model outputs to a temp folder
],
"console": "integratedTerminal", // show output in the integrated terminal
"cwd": "${workspaceFolder}/devtools", // set working directory to devtools from the root of the project
"justMyCode": true, // step through only axolotl code
"env": {"CUDA_VISIBLE_DEVICES": "0", // Since we aren't doing distributed training, we need to limit to one GPU
"HF_HOME": "${workspaceFolder}/devtools/temp_debug/.hf-cache"}, // send HF cache to a temp folder
"preLaunchTask": "cleanup-for-dataprep", // delete temp folders (see below)
}
]
}</code></pre>
<div class="sourceCode" id="cb4"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="er">//</span> <span class="er">.vscode/launch.json</span></span>
<span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb4-3"><a href="#cb4-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"version"</span><span class="fu">:</span> <span class="st">"0.2.0"</span><span class="fu">,</span></span>
<span id="cb4-4"><a href="#cb4-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"configurations"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb4-5"><a href="#cb4-5" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span></span>
<span id="cb4-6"><a href="#cb4-6" aria-hidden="true" tabindex="-1"></a> <span class="dt">"name"</span><span class="fu">:</span> <span class="st">"Debug axolotl prompt - chat_template"</span><span class="fu">,</span></span>
<span id="cb4-7"><a href="#cb4-7" aria-hidden="true" tabindex="-1"></a> <span class="dt">"type"</span><span class="fu">:</span> <span class="st">"python"</span><span class="fu">,</span></span>
<span id="cb4-8"><a href="#cb4-8" aria-hidden="true" tabindex="-1"></a> <span class="dt">"module"</span><span class="fu">:</span> <span class="st">"accelerate.commands.launch"</span><span class="fu">,</span></span>
<span id="cb4-9"><a href="#cb4-9" aria-hidden="true" tabindex="-1"></a> <span class="dt">"request"</span><span class="fu">:</span> <span class="st">"launch"</span><span class="fu">,</span></span>
<span id="cb4-10"><a href="#cb4-10" aria-hidden="true" tabindex="-1"></a> <span class="dt">"args"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb4-11"><a href="#cb4-11" aria-hidden="true" tabindex="-1"></a> <span class="st">"-m"</span><span class="ot">,</span> <span class="st">"axolotl.cli.train"</span><span class="ot">,</span> <span class="st">"dev_chat_template.yml"</span><span class="ot">,</span></span>
<span id="cb4-12"><a href="#cb4-12" aria-hidden="true" tabindex="-1"></a> <span class="er">//</span> <span class="er">The</span> <span class="er">flags</span> <span class="er">below</span> <span class="er">simplify</span> <span class="er">debugging</span> <span class="er">by</span> <span class="er">overriding</span> <span class="er">the</span> <span class="er">axolotl</span> <span class="er">config</span></span>
<span id="cb4-13"><a href="#cb4-13" aria-hidden="true" tabindex="-1"></a> <span class="er">//</span> <span class="er">with</span> <span class="er">the</span> <span class="er">debugging</span> <span class="er">tips</span> <span class="er">above.</span> <span class="er">Modify</span> <span class="er">as</span> <span class="er">needed.</span></span>
<span id="cb4-14"><a href="#cb4-14" aria-hidden="true" tabindex="-1"></a> <span class="st">"--dataset_processes=1"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">limits</span> <span class="er">data</span> <span class="er">preprocessing</span> <span class="er">to</span> <span class="er">one</span> <span class="er">process</span></span>
<span id="cb4-15"><a href="#cb4-15" aria-hidden="true" tabindex="-1"></a> <span class="st">"--max_steps=1"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">limits</span> <span class="er">training</span> <span class="er">to</span> <span class="er">just</span> <span class="er">one</span> <span class="er">step</span></span>
<span id="cb4-16"><a href="#cb4-16" aria-hidden="true" tabindex="-1"></a> <span class="st">"--batch_size=1"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">minimizes</span> <span class="er">batch</span> <span class="er">size</span></span>
<span id="cb4-17"><a href="#cb4-17" aria-hidden="true" tabindex="-1"></a> <span class="st">"--micro_batch_size=1"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">minimizes</span> <span class="er">batch</span> <span class="er">size</span></span>
<span id="cb4-18"><a href="#cb4-18" aria-hidden="true" tabindex="-1"></a> <span class="st">"--val_set_size=0"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">disables</span> <span class="er">validation</span></span>
<span id="cb4-19"><a href="#cb4-19" aria-hidden="true" tabindex="-1"></a> <span class="st">"--sample_packing=False"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">disables</span> <span class="er">sample</span> <span class="er">packing</span> <span class="er">which</span> <span class="er">is</span> <span class="er">necessary</span> <span class="er">for</span> <span class="er">small</span> <span class="er">datasets</span></span>
<span id="cb4-20"><a href="#cb4-20" aria-hidden="true" tabindex="-1"></a> <span class="st">"--eval_sample_packing=False"</span><span class="ot">,</span><span class="er">//</span> <span class="er">disables</span> <span class="er">sample</span> <span class="er">packing</span> <span class="er">on</span> <span class="er">eval</span> <span class="er">set</span></span>
<span id="cb4-21"><a href="#cb4-21" aria-hidden="true" tabindex="-1"></a> <span class="st">"--dataset_prepared_path=temp_debug/axolotl_outputs/data"</span><span class="ot">,</span> <span class="er">//</span> <span class="er">send</span> <span class="er">data</span> <span class="er">outputs</span> <span class="er">to</span> <span class="er">a</span> <span class="er">temp</span> <span class="er">folder</span></span>
<span id="cb4-22"><a href="#cb4-22" aria-hidden="true" tabindex="-1"></a> <span class="st">"--output_dir=temp_debug/axolotl_outputs/model"</span> <span class="er">//</span> <span class="er">send</span> <span class="er">model</span> <span class="er">outputs</span> <span class="er">to</span> <span class="er">a</span> <span class="er">temp</span> <span class="er">folder</span></span>
<span id="cb4-23"><a href="#cb4-23" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span><span class="fu">,</span></span>
<span id="cb4-24"><a href="#cb4-24" aria-hidden="true" tabindex="-1"></a> <span class="dt">"console"</span><span class="fu">:</span> <span class="st">"integratedTerminal"</span><span class="fu">,</span> <span class="er">//</span> <span class="er">show</span> <span class="er">output</span> <span class="er">in</span> <span class="er">the</span> <span class="er">integrated</span> <span class="er">terminal</span></span>
<span id="cb4-25"><a href="#cb4-25" aria-hidden="true" tabindex="-1"></a> <span class="dt">"cwd"</span><span class="fu">:</span> <span class="st">"${workspaceFolder}/devtools"</span><span class="fu">,</span> <span class="er">//</span> <span class="er">set</span> <span class="er">working</span> <span class="er">directory</span> <span class="er">to</span> <span class="er">devtools</span> <span class="er">from</span> <span class="er">the</span> <span class="er">root</span> <span class="er">of</span> <span class="er">the</span> <span class="er">project</span></span>
<span id="cb4-26"><a href="#cb4-26" aria-hidden="true" tabindex="-1"></a> <span class="dt">"justMyCode"</span><span class="fu">:</span> <span class="kw">true</span><span class="fu">,</span> <span class="er">//</span> <span class="er">step</span> <span class="er">through</span> <span class="er">only</span> <span class="er">axolotl</span> <span class="er">code</span></span>
<span id="cb4-27"><a href="#cb4-27" aria-hidden="true" tabindex="-1"></a> <span class="dt">"env"</span><span class="fu">:</span> <span class="fu">{</span><span class="dt">"CUDA_VISIBLE_DEVICES"</span><span class="fu">:</span> <span class="st">"0"</span><span class="fu">,</span> <span class="er">//</span> <span class="er">Since</span> <span class="er">we</span> <span class="er">aren't</span> <span class="er">doing</span> <span class="er">distributed</span> <span class="er">training</span><span class="fu">,</span> <span class="er">we</span> <span class="er">need</span> <span class="er">to</span> <span class="er">limit</span> <span class="er">to</span> <span class="er">one</span> <span class="er">GPU</span></span>
<span id="cb4-28"><a href="#cb4-28" aria-hidden="true" tabindex="-1"></a> <span class="dt">"HF_HOME"</span><span class="fu">:</span> <span class="st">"${workspaceFolder}/devtools/temp_debug/.hf-cache"</span><span class="fu">},</span> <span class="er">//</span> <span class="er">send</span> <span class="er">HF</span> <span class="er">cache</span> <span class="er">to</span> <span class="er">a</span> <span class="er">temp</span> <span class="er">folder</span></span>
<span id="cb4-29"><a href="#cb4-29" aria-hidden="true" tabindex="-1"></a> <span class="dt">"preLaunchTask"</span><span class="fu">:</span> <span class="st">"cleanup-for-dataprep"</span><span class="fu">,</span> <span class="er">//</span> <span class="er">delete</span> <span class="er">temp</span> <span class="er">folders</span> <span class="er">(see</span> <span class="er">below)</span></span>
<span id="cb4-30"><a href="#cb4-30" aria-hidden="true" tabindex="-1"></a> <span class="fu">}</span></span>
<span id="cb4-31"><a href="#cb4-31" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb4-32"><a href="#cb4-32" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p><strong>Additional notes about this configuration:</strong></p>
<ul>
<li>The argument <code>justMyCode</code> is set to <code>true</code> such that you step through only the axolotl code. If you want to step into dependencies, set this to <code>false</code>.</li>
@@ -498,34 +604,34 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<p>[!Tip] You may not want to delete these folders. For example, if you are debugging model training instead of data pre-processing, you may NOT want to delete the cache or output folders. You may also need to add additional tasks to the <code>tasks.json</code> file depending on your use case.</p>
</blockquote>
<p>Below is the <a href="../.vscode/tasks.json">./vscode/tasks.json</a> file that defines the <code>cleanup-for-dataprep</code> task. This task is run before each debugging session when you use the above configuration. Note how there are two tasks that delete the two folders mentioned above. The third task <code>cleanup-for-dataprep</code> is a composite task that combines the two tasks. A composite task is necessary because VSCode does not allow you to specify multiple tasks in the <code>preLaunchTask</code> argument of the <code>launch.json</code> file.</p>
<pre class="jsonc"><code>// .vscode/tasks.json
// this file is used by launch.json
{
"version": "2.0.0",
"tasks": [
// this task changes into the devtools directory and deletes the temp_debug/axolotl_outputs folder
{
"label": "delete-outputs",
"type": "shell",
"command": "rm -rf temp_debug/axolotl_outputs",
"options":{ "cwd": "${workspaceFolder}/devtools"},
"problemMatcher": []
},
// this task changes into the devtools directory and deletes the `temp_debug/.hf-cache/datasets` folder
{
"label": "delete-temp-hf-dataset-cache",
"type": "shell",
"command": "rm -rf temp_debug/.hf-cache/datasets",
"options":{ "cwd": "${workspaceFolder}/devtools"},
"problemMatcher": []
},
// this task combines the two tasks above
{
"label": "cleanup-for-dataprep",
"dependsOn": ["delete-outputs", "delete-temp-hf-dataset-cache"],
}
]
}</code></pre>
<div class="sourceCode" id="cb5"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="er">//</span> <span class="er">.vscode/tasks.json</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="er">//</span> <span class="er">this</span> <span class="er">file</span> <span class="er">is</span> <span class="er">used</span> <span class="er">by</span> <span class="er">launch.json</span></span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb5-4"><a href="#cb5-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"version"</span><span class="fu">:</span> <span class="st">"2.0.0"</span><span class="fu">,</span></span>
<span id="cb5-5"><a href="#cb5-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"tasks"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb5-6"><a href="#cb5-6" aria-hidden="true" tabindex="-1"></a> <span class="er">//</span> <span class="er">this</span> <span class="er">task</span> <span class="er">changes</span> <span class="er">into</span> <span class="er">the</span> <span class="er">devtools</span> <span class="er">directory</span> <span class="er">and</span> <span class="er">deletes</span> <span class="er">the</span> <span class="er">temp_debug/axolotl_outputs</span> <span class="er">folder</span></span>
<span id="cb5-7"><a href="#cb5-7" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span></span>
<span id="cb5-8"><a href="#cb5-8" aria-hidden="true" tabindex="-1"></a> <span class="dt">"label"</span><span class="fu">:</span> <span class="st">"delete-outputs"</span><span class="fu">,</span></span>
<span id="cb5-9"><a href="#cb5-9" aria-hidden="true" tabindex="-1"></a> <span class="dt">"type"</span><span class="fu">:</span> <span class="st">"shell"</span><span class="fu">,</span></span>
<span id="cb5-10"><a href="#cb5-10" aria-hidden="true" tabindex="-1"></a> <span class="dt">"command"</span><span class="fu">:</span> <span class="st">"rm -rf temp_debug/axolotl_outputs"</span><span class="fu">,</span></span>
<span id="cb5-11"><a href="#cb5-11" aria-hidden="true" tabindex="-1"></a> <span class="dt">"options"</span><span class="fu">:{</span> <span class="dt">"cwd"</span><span class="fu">:</span> <span class="st">"${workspaceFolder}/devtools"</span><span class="fu">},</span></span>
<span id="cb5-12"><a href="#cb5-12" aria-hidden="true" tabindex="-1"></a> <span class="dt">"problemMatcher"</span><span class="fu">:</span> <span class="ot">[]</span></span>
<span id="cb5-13"><a href="#cb5-13" aria-hidden="true" tabindex="-1"></a> <span class="fu">}</span><span class="ot">,</span></span>
<span id="cb5-14"><a href="#cb5-14" aria-hidden="true" tabindex="-1"></a> <span class="er">//</span> <span class="er">this</span> <span class="er">task</span> <span class="er">changes</span> <span class="er">into</span> <span class="er">the</span> <span class="er">devtools</span> <span class="er">directory</span> <span class="er">and</span> <span class="er">deletes</span> <span class="er">the</span> <span class="er">`temp_debug/.hf-cache/datasets`</span> <span class="er">folder</span></span>
<span id="cb5-15"><a href="#cb5-15" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span></span>
<span id="cb5-16"><a href="#cb5-16" aria-hidden="true" tabindex="-1"></a> <span class="dt">"label"</span><span class="fu">:</span> <span class="st">"delete-temp-hf-dataset-cache"</span><span class="fu">,</span></span>
<span id="cb5-17"><a href="#cb5-17" aria-hidden="true" tabindex="-1"></a> <span class="dt">"type"</span><span class="fu">:</span> <span class="st">"shell"</span><span class="fu">,</span></span>
<span id="cb5-18"><a href="#cb5-18" aria-hidden="true" tabindex="-1"></a> <span class="dt">"command"</span><span class="fu">:</span> <span class="st">"rm -rf temp_debug/.hf-cache/datasets"</span><span class="fu">,</span></span>
<span id="cb5-19"><a href="#cb5-19" aria-hidden="true" tabindex="-1"></a> <span class="dt">"options"</span><span class="fu">:{</span> <span class="dt">"cwd"</span><span class="fu">:</span> <span class="st">"${workspaceFolder}/devtools"</span><span class="fu">},</span></span>
<span id="cb5-20"><a href="#cb5-20" aria-hidden="true" tabindex="-1"></a> <span class="dt">"problemMatcher"</span><span class="fu">:</span> <span class="ot">[]</span></span>
<span id="cb5-21"><a href="#cb5-21" aria-hidden="true" tabindex="-1"></a> <span class="fu">}</span><span class="ot">,</span></span>
<span id="cb5-22"><a href="#cb5-22" aria-hidden="true" tabindex="-1"></a> <span class="er">//</span> <span class="er">this</span> <span class="er">task</span> <span class="er">combines</span> <span class="er">the</span> <span class="er">two</span> <span class="er">tasks</span> <span class="er">above</span></span>
<span id="cb5-23"><a href="#cb5-23" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span></span>
<span id="cb5-24"><a href="#cb5-24" aria-hidden="true" tabindex="-1"></a> <span class="dt">"label"</span><span class="fu">:</span> <span class="st">"cleanup-for-dataprep"</span><span class="fu">,</span></span>
<span id="cb5-25"><a href="#cb5-25" aria-hidden="true" tabindex="-1"></a> <span class="dt">"dependsOn"</span><span class="fu">:</span> <span class="ot">[</span><span class="st">"delete-outputs"</span><span class="ot">,</span> <span class="st">"delete-temp-hf-dataset-cache"</span><span class="ot">]</span><span class="fu">,</span></span>
<span id="cb5-26"><a href="#cb5-26" aria-hidden="true" tabindex="-1"></a> <span class="fu">}</span></span>
<span id="cb5-27"><a href="#cb5-27" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb5-28"><a href="#cb5-28" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="customizing-your-debugger" class="level3">
<h3 class="anchored" data-anchor-id="customizing-your-debugger">Customizing your debugger</h3>
@@ -549,19 +655,19 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<section id="setup-1" class="level3">
<h3 class="anchored" data-anchor-id="setup-1">Setup</h3>
<p>On the host that is running axolotl (ex: if you are using a remote host), clone the axolotl repo and change your current directory to the root:</p>
<div class="sourceCode" id="cb5"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="fu">git</span> clone https://github.com/axolotl-ai-cloud/axolotl</span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="bu">cd</span> axolotl</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="fu">git</span> clone https://github.com/axolotl-ai-cloud/axolotl</span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a><span class="bu">cd</span> axolotl</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<blockquote class="blockquote">
<p>[!Tip] If you already have axolotl cloned on your host, make sure you have the latest changes and change into the root of the project.</p>
</blockquote>
<p>Next, run the desired docker image and mount the current directory. Below is a docker command you can run to do this:<a href="#fn2" class="footnote-ref" id="fnref2" role="doc-noteref"><sup>2</sup></a></p>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="ex">docker</span> run <span class="at">--privileged</span> <span class="at">--gpus</span> <span class="st">'"all"'</span> <span class="at">--shm-size</span> 10g <span class="at">--rm</span> <span class="at">-it</span> <span class="at">--name</span> axolotl <span class="at">--ipc</span><span class="op">=</span>host <span class="at">--ulimit</span> memlock=-1 <span class="at">--ulimit</span> stack=67108864 <span class="at">--mount</span> type=bind,src=<span class="st">"</span><span class="va">${PWD}</span><span class="st">"</span>,target=/workspace/axolotl <span class="at">-v</span> <span class="va">${HOME}</span>/.cache/huggingface:/root/.cache/huggingface axolotlai/axolotl:main-py3.10-cu118-2.0.1</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb7"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="ex">docker</span> run <span class="at">--privileged</span> <span class="at">--gpus</span> <span class="st">'"all"'</span> <span class="at">--shm-size</span> 10g <span class="at">--rm</span> <span class="at">-it</span> <span class="at">--name</span> axolotl <span class="at">--ipc</span><span class="op">=</span>host <span class="at">--ulimit</span> memlock=-1 <span class="at">--ulimit</span> stack=67108864 <span class="at">--mount</span> type=bind,src=<span class="st">"</span><span class="va">${PWD}</span><span class="st">"</span>,target=/workspace/axolotl <span class="at">-v</span> <span class="va">${HOME}</span>/.cache/huggingface:/root/.cache/huggingface axolotlai/axolotl:main-py3.10-cu118-2.0.1</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<blockquote class="blockquote">
<p>[!Tip] To understand which containers are available, see the <a href="../README.md#docker">Docker section of the README</a> and the <a href="https://hub.docker.com/r/axolotlai/axolotl/tags">DockerHub repo</a>. For details of how the Docker containers are built, see axolotls <a href="../.github/workflows/main.yml">Docker CI builds</a>.</p>
</blockquote>
<p>You will now be in the container. Next, perform an editable install of Axolotl:</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install packaging</span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install <span class="at">--no-build-isolation</span> <span class="at">-e</span> <span class="st">'.[flash-attn,deepspeed]'</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb8"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install packaging</span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="ex">pip3</span> install <span class="at">--no-build-isolation</span> <span class="at">-e</span> <span class="st">'.[flash-attn,deepspeed]'</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="attach-to-container" class="level3">
<h3 class="anchored" data-anchor-id="attach-to-container">Attach To Container</h3>

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">FAQ</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/faq.html">FAQ</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,30 +207,206 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
<div id="quarto-sidebar-glass" class="quarto-sidebar-collapse-item" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item"></div>
<!-- margin-sidebar -->
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar zindex-bottom">
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
<nav id="TOC" role="doc-toc" class="toc-active">
<h2 id="toc-title">On this page</h2>
<ul>
<li><a href="#general" id="toc-general" class="nav-link active" data-scroll-target="#general">General</a></li>
<li><a href="#chat-templates" id="toc-chat-templates" class="nav-link" data-scroll-target="#chat-templates">Chat templates</a></li>
</ul>
</nav>
</div>
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/faq.html">FAQ</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">FAQ</h1>
</div>
@@ -320,6 +430,8 @@ ul.task-list li input[type="checkbox"] {
</header>
<section id="general" class="level3">
<h3 class="anchored" data-anchor-id="general">General</h3>
<p><strong>Q: The trainer stopped and hasnt progressed in several minutes.</strong></p>
<blockquote class="blockquote">
<p>A: Usually an issue with the GPUs communicating with each other. See the <a href="../docs/nccl.html">NCCL doc</a></p>
@@ -340,12 +452,32 @@ ul.task-list li input[type="checkbox"] {
<blockquote class="blockquote">
<p>A: This is usually an issue with the GPU. This can be resolved through setting the os environment variable <code>CUDA_VISIBLE_DEVICES=0</code>. If you are on runpod, this is usually a pod issue. Starting a new pod should take care of it.</p>
</blockquote>
</section>
<section id="chat-templates" class="level3">
<h3 class="anchored" data-anchor-id="chat-templates">Chat templates</h3>
<p><strong>Q: <code>jinja2.exceptions.UndefinedError: 'dict object' has no attribute 'content' / 'role' / ____</code></strong></p>
<blockquote class="blockquote">
<p>A: This means that the property mapping for the stated attribute does not exist when building <code>chat_template</code> prompt. For example, if <code>no attribute 'content'</code>, please check you have added the correct mapping for <code>content</code> under <code>message_property_mappings</code>.</p>
</blockquote>
<p><strong>Q: <code>Empty template generated for turn ___</code></strong></p>
<blockquote class="blockquote">
<p>A: The <code>content</code> is empty for that turn.</p>
</blockquote>
<p><strong>Q: <code>Could not find content start/end boundary for turn __</code></strong></p>
<blockquote class="blockquote">
<p>A: The specific turns start/end could not be detected. Please ensure you have set the <code>eos_token</code> following your <code>chat_template</code>. Otherwise, this could be a <code>chat_template</code> which doesnt use proper boundaries for each turn (like system). On the rare occurrence, make sure your content is not <code>[[dummy_message]]</code>. Please let us know about this.</p>
</blockquote>
<p><strong>Q: <code>Content end boundary is before start boundary for turn ___</code></strong></p>
<blockquote class="blockquote">
<p>A: This is an edge case which should not occur. Please create an Issue if this happens.</p>
</blockquote>
<p><strong>Q: <code>Content end boundary is the same as start boundary for turn ___. This is likely an empty turn.</code></strong></p>
<blockquote class="blockquote">
<p>A: This is likely an empty turn.</p>
</blockquote>
</section>
</main> <!-- /main -->
<script id="quarto-html-after-body" type="application/javascript">

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">FDSP + QLoRA</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">FDSP + QLoRA</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +207,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -305,7 +408,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">FDSP + QLoRA</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">FDSP + QLoRA</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">FDSP + QLoRA</h1>
</div>

View File

@@ -7,7 +7,7 @@
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<title>Getting Started with Axolotl Axolotl</title>
<title>Quickstart Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started with Axolotl</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/getting-started.html">Quickstart</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,18 +240,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -348,9 +451,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started with Axolotl</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/getting-started.html">Quickstart</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Getting Started with Axolotl</h1>
<h1 class="title">Quickstart</h1>
</div>
@@ -374,11 +477,11 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ol type="1">
<li>Download example configs:</li>
</ol>
<pre class="shell"><code>axolotl fetch examples</code></pre>
<div class="sourceCode" id="cb1"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> fetch examples</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<ol start="2" type="1">
<li>Run the training:</li>
</ol>
<pre class="shell"><code>axolotl train examples/llama-3/lora-1b.yml</code></pre>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train examples/llama-3/lora-1b.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Thats it! Lets understand what just happened.</p>
</section>
<section id="sec-understanding" class="level2" data-number="2">
@@ -446,24 +549,24 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ol start="3" type="1">
<li>Run the training:</li>
</ol>
<pre class="shell"><code>axolotl train my_training.yml</code></pre>
<div class="sourceCode" id="cb7"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train my_training.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="sec-common-tasks" class="level2" data-number="4">
<h2 data-number="4" class="anchored" data-anchor-id="sec-common-tasks"><span class="header-section-number">4</span> Common Tasks</h2>
<section id="sec-testing" class="level3" data-number="4.1">
<h3 data-number="4.1" class="anchored" data-anchor-id="sec-testing"><span class="header-section-number">4.1</span> Testing Your Model</h3>
<p>After training, test your model:</p>
<pre class="shell"><code>axolotl inference my_training.yml --lora-model-dir="./outputs/lora-out"</code></pre>
<div class="sourceCode" id="cb8"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference my_training.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="sec-preprocessing" class="level3" data-number="4.2">
<h3 data-number="4.2" class="anchored" data-anchor-id="sec-preprocessing"><span class="header-section-number">4.2</span> Preprocessing Data</h3>
<p>For large datasets, preprocess first:</p>
<pre class="shell"><code>axolotl preprocess my_training.yml</code></pre>
<div class="sourceCode" id="cb9"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> preprocess my_training.yml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="sec-ui" class="level3" data-number="4.3">
<h3 data-number="4.3" class="anchored" data-anchor-id="sec-ui"><span class="header-section-number">4.3</span> Using a UI</h3>
<p>Launch a Gradio interface:</p>
<pre class="shell"><code>axolotl inference my_training.yml --lora-model-dir="./outputs/lora-out" --gradio</code></pre>
<div class="sourceCode" id="cb10"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> inference my_training.yml <span class="at">--lora-model-dir</span><span class="op">=</span><span class="st">"./outputs/lora-out"</span> <span class="at">--gradio</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
<section id="sec-next-steps" class="level2" data-number="5">

View File

@@ -7,7 +7,7 @@
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<title>Inference Guide Axolotl</title>
<title>Inference Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/inference.html">Inference Guide</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/inference.html">Inference</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,18 +240,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -357,9 +460,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/inference.html">Inference Guide</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/inference.html">Inference</a></li></ol></nav>
<div class="quarto-title">
<div class="quarto-title-block"><div><h1 class="title">Inference Guide</h1><button type="button" class="btn code-tools-button" id="quarto-code-tools-source"><i class="bi"></i> Code</button></div></div>
<h1 class="title">Inference</h1>
</div>
@@ -505,8 +608,6 @@ Warning
<p>For more details, see our <a href="../docs/debugging.html">debugging guide</a>.</p>
<!-- -->
</section>
</section>
@@ -598,57 +699,6 @@ window.document.addEventListener("DOMContentLoaded", function (event) {
container: window.document.getElementById('quarto-embedded-source-code-modal')
});
clipboardModal.on('success', onCopySuccess);
}
const viewSource = window.document.getElementById('quarto-view-source') ||
window.document.getElementById('quarto-code-tools-source');
if (viewSource) {
const sourceUrl = viewSource.getAttribute("data-quarto-source-url");
viewSource.addEventListener("click", function(e) {
if (sourceUrl) {
// rstudio viewer pane
if (/\bcapabilities=\b/.test(window.location)) {
window.open(sourceUrl);
} else {
window.location.href = sourceUrl;
}
} else {
const modal = new bootstrap.Modal(document.getElementById('quarto-embedded-source-code-modal'));
modal.show();
}
return false;
});
}
function toggleCodeHandler(show) {
return function(e) {
const detailsSrc = window.document.querySelectorAll(".cell > details > .sourceCode");
for (let i=0; i<detailsSrc.length; i++) {
const details = detailsSrc[i].parentElement;
if (show) {
details.open = true;
} else {
details.removeAttribute("open");
}
}
const cellCodeDivs = window.document.querySelectorAll(".cell > .sourceCode");
const fromCls = show ? "hidden" : "unhidden";
const toCls = show ? "unhidden" : "hidden";
for (let i=0; i<cellCodeDivs.length; i++) {
const codeDiv = cellCodeDivs[i];
if (codeDiv.classList.contains(fromCls)) {
codeDiv.classList.remove(fromCls);
codeDiv.classList.add(toCls);
}
}
return false;
}
}
const hideAllCode = window.document.getElementById("quarto-hide-all-code");
if (hideAllCode) {
hideAllCode.addEventListener("click", toggleCodeHandler(false));
}
const showAllCode = window.document.getElementById("quarto-show-all-code");
if (showAllCode) {
showAllCode.addEventListener("click", toggleCodeHandler(true));
}
var localhostRegex = new RegExp(/^(?:http|https):\/\/localhost\:?[0-9]*\//);
var mailtoRegex = new RegExp(/^mailto:/);
@@ -978,156 +1028,7 @@ window.document.addEventListener("DOMContentLoaded", function (event) {
}
}
});
</script><div class="modal fade" id="quarto-embedded-source-code-modal" tabindex="-1" aria-labelledby="quarto-embedded-source-code-modal-label" aria-hidden="true"><div class="modal-dialog modal-dialog-scrollable"><div class="modal-content"><div class="modal-header"><h5 class="modal-title" id="quarto-embedded-source-code-modal-label">Source Code</h5><button class="btn-close" data-bs-dismiss="modal"></button></div><div class="modal-body"><div class="">
<div class="sourceCode" id="cb11" data-shortcodes="false"><pre class="sourceCode markdown code-with-copy"><code class="sourceCode markdown"><span id="cb11-1"><a href="#cb11-1" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb11-2"><a href="#cb11-2" aria-hidden="true" tabindex="-1"></a><span class="an">title:</span><span class="co"> "Inference Guide"</span></span>
<span id="cb11-3"><a href="#cb11-3" aria-hidden="true" tabindex="-1"></a><span class="an">format:</span></span>
<span id="cb11-4"><a href="#cb11-4" aria-hidden="true" tabindex="-1"></a><span class="co"> html:</span></span>
<span id="cb11-5"><a href="#cb11-5" aria-hidden="true" tabindex="-1"></a><span class="co"> toc: true</span></span>
<span id="cb11-6"><a href="#cb11-6" aria-hidden="true" tabindex="-1"></a><span class="co"> toc-depth: 3</span></span>
<span id="cb11-7"><a href="#cb11-7" aria-hidden="true" tabindex="-1"></a><span class="co"> number-sections: true</span></span>
<span id="cb11-8"><a href="#cb11-8" aria-hidden="true" tabindex="-1"></a><span class="co"> code-tools: true</span></span>
<span id="cb11-9"><a href="#cb11-9" aria-hidden="true" tabindex="-1"></a><span class="an">execute:</span></span>
<span id="cb11-10"><a href="#cb11-10" aria-hidden="true" tabindex="-1"></a><span class="co"> enabled: false</span></span>
<span id="cb11-11"><a href="#cb11-11" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb11-12"><a href="#cb11-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-13"><a href="#cb11-13" aria-hidden="true" tabindex="-1"></a>This guide covers how to use your trained models for inference, including model loading, interactive testing, and common troubleshooting steps.</span>
<span id="cb11-14"><a href="#cb11-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-15"><a href="#cb11-15" aria-hidden="true" tabindex="-1"></a><span class="fu">## Quick Start {#sec-quickstart}</span></span>
<span id="cb11-16"><a href="#cb11-16" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-17"><a href="#cb11-17" aria-hidden="true" tabindex="-1"></a><span class="fu">### Basic Inference {#sec-basic}</span></span>
<span id="cb11-18"><a href="#cb11-18" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-19"><a href="#cb11-19" aria-hidden="true" tabindex="-1"></a>::: {.panel-tabset}</span>
<span id="cb11-20"><a href="#cb11-20" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-21"><a href="#cb11-21" aria-hidden="true" tabindex="-1"></a><span class="fu">## LoRA Models</span></span>
<span id="cb11-22"><a href="#cb11-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-23"><a href="#cb11-23" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-24"><a href="#cb11-24" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl inference your_config.yml --lora-model-dir="./lora-output-dir"</span></span>
<span id="cb11-25"><a href="#cb11-25" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-26"><a href="#cb11-26" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-27"><a href="#cb11-27" aria-hidden="true" tabindex="-1"></a><span class="fu">## Full Fine-tuned Models</span></span>
<span id="cb11-28"><a href="#cb11-28" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-29"><a href="#cb11-29" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-30"><a href="#cb11-30" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl inference your_config.yml --base-model="./completed-model"</span></span>
<span id="cb11-31"><a href="#cb11-31" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-32"><a href="#cb11-32" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-33"><a href="#cb11-33" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb11-34"><a href="#cb11-34" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-35"><a href="#cb11-35" aria-hidden="true" tabindex="-1"></a><span class="fu">## Advanced Usage {#sec-advanced}</span></span>
<span id="cb11-36"><a href="#cb11-36" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-37"><a href="#cb11-37" aria-hidden="true" tabindex="-1"></a><span class="fu">### Gradio Interface {#sec-gradio}</span></span>
<span id="cb11-38"><a href="#cb11-38" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-39"><a href="#cb11-39" aria-hidden="true" tabindex="-1"></a>Launch an interactive web interface:</span>
<span id="cb11-40"><a href="#cb11-40" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-41"><a href="#cb11-41" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-42"><a href="#cb11-42" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl inference your_config.yml --gradio</span></span>
<span id="cb11-43"><a href="#cb11-43" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-44"><a href="#cb11-44" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-45"><a href="#cb11-45" aria-hidden="true" tabindex="-1"></a><span class="fu">### File-based Prompts {#sec-file-prompts}</span></span>
<span id="cb11-46"><a href="#cb11-46" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-47"><a href="#cb11-47" aria-hidden="true" tabindex="-1"></a>Process prompts from a text file:</span>
<span id="cb11-48"><a href="#cb11-48" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-49"><a href="#cb11-49" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-50"><a href="#cb11-50" aria-hidden="true" tabindex="-1"></a><span class="in">cat /tmp/prompt.txt | axolotl inference your_config.yml \</span></span>
<span id="cb11-51"><a href="#cb11-51" aria-hidden="true" tabindex="-1"></a><span class="in"> --base-model="./completed-model" --prompter=None</span></span>
<span id="cb11-52"><a href="#cb11-52" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-53"><a href="#cb11-53" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-54"><a href="#cb11-54" aria-hidden="true" tabindex="-1"></a><span class="fu">### Memory Optimization {#sec-memory}</span></span>
<span id="cb11-55"><a href="#cb11-55" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-56"><a href="#cb11-56" aria-hidden="true" tabindex="-1"></a>For large models or limited memory:</span>
<span id="cb11-57"><a href="#cb11-57" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-58"><a href="#cb11-58" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-59"><a href="#cb11-59" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl inference your_config.yml --load-in-8bit=True</span></span>
<span id="cb11-60"><a href="#cb11-60" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-61"><a href="#cb11-61" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-62"><a href="#cb11-62" aria-hidden="true" tabindex="-1"></a><span class="fu">## Merging LoRA Weights {#sec-merging}</span></span>
<span id="cb11-63"><a href="#cb11-63" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-64"><a href="#cb11-64" aria-hidden="true" tabindex="-1"></a>Merge LoRA adapters with the base model:</span>
<span id="cb11-65"><a href="#cb11-65" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-66"><a href="#cb11-66" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-67"><a href="#cb11-67" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl merge-lora your_config.yml --lora-model-dir="./completed-model"</span></span>
<span id="cb11-68"><a href="#cb11-68" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-69"><a href="#cb11-69" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-70"><a href="#cb11-70" aria-hidden="true" tabindex="-1"></a><span class="fu">### Memory Management for Merging {#sec-memory-management}</span></span>
<span id="cb11-71"><a href="#cb11-71" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-72"><a href="#cb11-72" aria-hidden="true" tabindex="-1"></a>::: {.panel-tabset}</span>
<span id="cb11-73"><a href="#cb11-73" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-74"><a href="#cb11-74" aria-hidden="true" tabindex="-1"></a><span class="fu">## Configuration Options</span></span>
<span id="cb11-75"><a href="#cb11-75" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-76"><a href="#cb11-76" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb11-77"><a href="#cb11-77" aria-hidden="true" tabindex="-1"></a><span class="in">gpu_memory_limit: 20GiB # Adjust based on your GPU</span></span>
<span id="cb11-78"><a href="#cb11-78" aria-hidden="true" tabindex="-1"></a><span class="in">lora_on_cpu: true # Process on CPU if needed</span></span>
<span id="cb11-79"><a href="#cb11-79" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-80"><a href="#cb11-80" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-81"><a href="#cb11-81" aria-hidden="true" tabindex="-1"></a><span class="fu">## Force CPU Merging</span></span>
<span id="cb11-82"><a href="#cb11-82" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-83"><a href="#cb11-83" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-84"><a href="#cb11-84" aria-hidden="true" tabindex="-1"></a><span class="in">CUDA_VISIBLE_DEVICES="" axolotl merge-lora ...</span></span>
<span id="cb11-85"><a href="#cb11-85" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-86"><a href="#cb11-86" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-87"><a href="#cb11-87" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb11-88"><a href="#cb11-88" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-89"><a href="#cb11-89" aria-hidden="true" tabindex="-1"></a><span class="fu">## Tokenization {#sec-tokenization}</span></span>
<span id="cb11-90"><a href="#cb11-90" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-91"><a href="#cb11-91" aria-hidden="true" tabindex="-1"></a><span class="fu">### Common Issues {#sec-tokenization-issues}</span></span>
<span id="cb11-92"><a href="#cb11-92" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-93"><a href="#cb11-93" aria-hidden="true" tabindex="-1"></a>::: {.callout-warning}</span>
<span id="cb11-94"><a href="#cb11-94" aria-hidden="true" tabindex="-1"></a>Tokenization mismatches between training and inference are a common source of problems.</span>
<span id="cb11-95"><a href="#cb11-95" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb11-96"><a href="#cb11-96" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-97"><a href="#cb11-97" aria-hidden="true" tabindex="-1"></a>To debug:</span>
<span id="cb11-98"><a href="#cb11-98" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-99"><a href="#cb11-99" aria-hidden="true" tabindex="-1"></a><span class="ss">1. </span>Check training tokenization:</span>
<span id="cb11-100"><a href="#cb11-100" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb11-101"><a href="#cb11-101" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl preprocess your_config.yml --debug</span></span>
<span id="cb11-102"><a href="#cb11-102" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-103"><a href="#cb11-103" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-104"><a href="#cb11-104" aria-hidden="true" tabindex="-1"></a><span class="ss">2. </span>Verify inference tokenization by decoding tokens before model input</span>
<span id="cb11-105"><a href="#cb11-105" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-106"><a href="#cb11-106" aria-hidden="true" tabindex="-1"></a><span class="ss">3. </span>Compare token IDs between training and inference</span>
<span id="cb11-107"><a href="#cb11-107" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-108"><a href="#cb11-108" aria-hidden="true" tabindex="-1"></a><span class="fu">### Special Tokens {#sec-special-tokens}</span></span>
<span id="cb11-109"><a href="#cb11-109" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-110"><a href="#cb11-110" aria-hidden="true" tabindex="-1"></a>Configure special tokens in your YAML:</span>
<span id="cb11-111"><a href="#cb11-111" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-112"><a href="#cb11-112" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb11-113"><a href="#cb11-113" aria-hidden="true" tabindex="-1"></a><span class="in">special_tokens:</span></span>
<span id="cb11-114"><a href="#cb11-114" aria-hidden="true" tabindex="-1"></a><span class="in"> bos_token: "&lt;s&gt;"</span></span>
<span id="cb11-115"><a href="#cb11-115" aria-hidden="true" tabindex="-1"></a><span class="in"> eos_token: "&lt;/s&gt;"</span></span>
<span id="cb11-116"><a href="#cb11-116" aria-hidden="true" tabindex="-1"></a><span class="in"> unk_token: "&lt;unk&gt;"</span></span>
<span id="cb11-117"><a href="#cb11-117" aria-hidden="true" tabindex="-1"></a><span class="in">tokens:</span></span>
<span id="cb11-118"><a href="#cb11-118" aria-hidden="true" tabindex="-1"></a><span class="in"> - "&lt;|im_start|&gt;"</span></span>
<span id="cb11-119"><a href="#cb11-119" aria-hidden="true" tabindex="-1"></a><span class="in"> - "&lt;|im_end|&gt;"</span></span>
<span id="cb11-120"><a href="#cb11-120" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb11-121"><a href="#cb11-121" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-122"><a href="#cb11-122" aria-hidden="true" tabindex="-1"></a><span class="fu">## Troubleshooting {#sec-troubleshooting}</span></span>
<span id="cb11-123"><a href="#cb11-123" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-124"><a href="#cb11-124" aria-hidden="true" tabindex="-1"></a><span class="fu">### Common Problems {#sec-common-problems}</span></span>
<span id="cb11-125"><a href="#cb11-125" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-126"><a href="#cb11-126" aria-hidden="true" tabindex="-1"></a>::: {.panel-tabset}</span>
<span id="cb11-127"><a href="#cb11-127" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-128"><a href="#cb11-128" aria-hidden="true" tabindex="-1"></a><span class="fu">## Memory Issues</span></span>
<span id="cb11-129"><a href="#cb11-129" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-130"><a href="#cb11-130" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Use 8-bit loading</span>
<span id="cb11-131"><a href="#cb11-131" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Reduce batch sizes</span>
<span id="cb11-132"><a href="#cb11-132" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Try CPU offloading</span>
<span id="cb11-133"><a href="#cb11-133" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-134"><a href="#cb11-134" aria-hidden="true" tabindex="-1"></a><span class="fu">## Token Issues</span></span>
<span id="cb11-135"><a href="#cb11-135" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-136"><a href="#cb11-136" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Verify special tokens</span>
<span id="cb11-137"><a href="#cb11-137" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Check tokenizer settings</span>
<span id="cb11-138"><a href="#cb11-138" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Compare training and inference preprocessing</span>
<span id="cb11-139"><a href="#cb11-139" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-140"><a href="#cb11-140" aria-hidden="true" tabindex="-1"></a><span class="fu">## Performance Issues</span></span>
<span id="cb11-141"><a href="#cb11-141" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-142"><a href="#cb11-142" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Verify model loading</span>
<span id="cb11-143"><a href="#cb11-143" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Check prompt formatting</span>
<span id="cb11-144"><a href="#cb11-144" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Ensure temperature/sampling settings</span>
<span id="cb11-145"><a href="#cb11-145" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-146"><a href="#cb11-146" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb11-147"><a href="#cb11-147" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb11-148"><a href="#cb11-148" aria-hidden="true" tabindex="-1"></a>For more details, see our <span class="co">[</span><span class="ot">debugging guide</span><span class="co">](debugging.qmd)</span>.</span></code><button title="Copy to Clipboard" class="code-copy-button" data-in-quarto-modal=""><i class="bi"></i></button></pre></div>
</div></div></div></div></div>
</script>
</div> <!-- /content -->

View File

@@ -21,40 +21,6 @@ ul.task-list li input[type="checkbox"] {
margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */
vertical-align: middle;
}
/* CSS for syntax highlighting */
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
div.sourceCode { margin: 1em 0; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { display: inline-block; text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
{ counter-reset: source-line 0; }
pre.numberSource code > span
{ position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
{ content: counter(source-line);
position: relative; left: -1em; text-align: right; vertical-align: baseline;
border: none; display: inline-block;
-webkit-touch-callout: none; -webkit-user-select: none;
-khtml-user-select: none; -moz-user-select: none;
-ms-user-select: none; user-select: none;
padding: 0 4px; width: 4em;
}
pre.numberSource { margin-left: 3em; padding-left: 4px; }
div.sourceCode
{ }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
</style>
@@ -70,10 +36,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +79,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +96,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/input_output.html">Template-free prompt construction</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">Template-free prompt construction</li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +117,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +126,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +199,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,47 +207,199 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
<div id="quarto-sidebar-glass" class="quarto-sidebar-collapse-item" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item"></div>
<!-- margin-sidebar -->
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
<nav id="TOC" role="doc-toc" class="toc-active">
<h2 id="toc-title">On this page</h2>
<ul>
<li><a href="#background" id="toc-background" class="nav-link active" data-scroll-target="#background">Background</a>
<ul class="collapse">
<li><a href="#masking-inputs" id="toc-masking-inputs" class="nav-link" data-scroll-target="#masking-inputs">Masking Inputs</a></li>
<li><a href="#you-may-not-want-prompt-templates" id="toc-you-may-not-want-prompt-templates" class="nav-link" data-scroll-target="#you-may-not-want-prompt-templates">You may not want prompt templates</a></li>
<li><a href="#the-input_output-format" id="toc-the-input_output-format" class="nav-link" data-scroll-target="#the-input_output-format">The <code>input_output</code> format</a></li>
</ul></li>
<li><a href="#usage" id="toc-usage" class="nav-link" data-scroll-target="#usage">Usage</a>
<ul class="collapse">
<li><a href="#prepare-data" id="toc-prepare-data" class="nav-link" data-scroll-target="#prepare-data">1. Prepare Data</a></li>
<li><a href="#use-type-input_output" id="toc-use-type-input_output" class="nav-link" data-scroll-target="#use-type-input_output">2. Use <code>type: input_output</code></a></li>
<li><a href="#check-the-prompts" id="toc-check-the-prompts" class="nav-link" data-scroll-target="#check-the-prompts">3. Check the prompts</a></li>
</ul></li>
</ul>
</nav>
<div id="quarto-margin-sidebar" class="sidebar margin-sidebar zindex-bottom">
</div>
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/input_output.html">Template-free prompt construction</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default">
<div class="quarto-title">
<h1 class="title">Template-free prompt construction</h1>
</div>
@@ -371,258 +423,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
</header>
<!-- TOC -->
<ul>
<li><a href="#background">Background</a>
<ul>
<li><a href="#masking-inputs">Masking Inputs</a></li>
<li><a href="#you-may-not-want-prompt-templates">You may not want prompt templates</a></li>
<li><a href="#the-input_output-format">The <code>input_output</code> format</a></li>
</ul></li>
<li><a href="#usage">Usage</a>
<ul>
<li><a href="#1-prepare-data">1. Prepare Data</a></li>
<li><a href="#2-use-type-input_output">2. Use <code>type: input_output</code></a></li>
<li><a href="#3-check-the-prompts">3. Check the prompts</a></li>
</ul></li>
</ul>
<!-- /TOC -->
<p><a id="markdown-background" name="background"></a></p>
<section id="background" class="level2">
<h2 class="anchored" data-anchor-id="background">Background</h2>
<p><a id="markdown-masking-inputs" name="masking-inputs"></a></p>
<section id="masking-inputs" class="level3">
<h3 class="anchored" data-anchor-id="masking-inputs">Masking Inputs</h3>
<p>One of the most popular features of <a href="https://github.com/axolotl-ai-cloud/axolotl">axolotl</a> is setting the following configuration value:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>If you declare a <a href="https://github.com/axolotl-ai-cloud/axolotl?tab=readme-ov-file#dataset">dataset formats</a> such as <code>alpaca</code> or <code>chatml</code>, axolotl knows what is an input (i.e.&nbsp;human) vs.&nbsp;an output (i.e.&nbsp;the assistant) and masks the input labels so that your model can focus on predicting the outputs only.</p>
<p><a id="markdown-you-may-not-want-prompt-templates" name="you-may-not-want-prompt-templates"></a></p>
</section>
<section id="you-may-not-want-prompt-templates" class="level3">
<h3 class="anchored" data-anchor-id="you-may-not-want-prompt-templates">You may not want prompt templates</h3>
<p>However, there are many situations where you dont want to use one of these formats or templates. This is because they can:</p>
<ul>
<li>Add unnecessary boilerplate to your prompts.</li>
<li>Create artifacts like special delimiters <code>&lt;|im_start|&gt;</code> that can quickly become footguns if you dont include them correctly at inference time.</li>
<li>Enforce a <em>chat</em> interface when you do not want one. Sometimes you just want to fine-tune a model to a very specific task and do NOT want multi-turn conversations, roles, etc.</li>
<li>Limit you to only certain roles that the template allows.</li>
</ul>
<p><a id="markdown-the-inputoutput-format" name="the-inputoutput-format"></a></p>
</section>
<section id="the-input_output-format" class="level3">
<h3 class="anchored" data-anchor-id="the-input_output-format">The <code>input_output</code> format</h3>
<p>You can construct your prompts without a template by using the <code>input_output</code> format, by setting <code>type: input_output</code> in your configuration file like this:</p>
<p><strong>config.yml</strong></p>
<div class="sourceCode" id="cb2"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span><span class="co"> # Mask segments of your data</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> output.jsonl</span></span>
<span id="cb2-4"><a href="#cb2-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> input_output</span><span class="co"> # use template free prompt construction</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Unlike <code>type: completion</code>, which is also template-free, <code>type: input_output</code> allows you to mask segments of your text. More details on how this works are described below.</p>
<p><a id="markdown-usage" name="usage"></a></p>
</section>
</section>
<section id="usage" class="level2">
<h2 class="anchored" data-anchor-id="usage">Usage</h2>
<p>This is how you can use the <code>input_output</code> format:</p>
<p><a id="markdown-1-prepare-data" name="1-prepare-data"></a></p>
<section id="prepare-data" class="level3">
<h3 class="anchored" data-anchor-id="prepare-data">1. Prepare Data</h3>
<p>To use the <code>input_output</code> format, collect your data in the following format into a jsonl file (below is the first row from the file <code>output</code>.jsonl` pretty printed):</p>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="ex">$</span> head <span class="at">-n1</span> output.jsonl <span class="kw">|</span> <span class="ex">python</span> <span class="at">-m</span> json.tool</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="cell-output cell-output-stdout">
<pre><code>{
"segments": [
{
"label": true,
"text": "&lt;s&gt;Hello\n"
},
{
"label": true,
"text": "hi there!. "
},
{
"label": false,
"text": "goodbye "
},
{
"label": true,
"text": "farewell&lt;/s&gt;"
}
]
}</code></pre>
</div>
<p>Set <code>label:false</code> when you want to mask a segment of text so that the model isnt trained on it. Some things to keep in mind:</p>
<blockquote class="blockquote">
<p>[!IMPORTANT] 1. <strong>EOS, BOS, spaces, newlines etc. are entirely up to you. Axolotl concatenates all the segments as-is.</strong> The tokenizer doesnt add anything additional. Notice how I added spaces, newlines, <code>&lt;s&gt;</code> (BOS), and <code>&lt;/s&gt;</code> (EOS) myself. 2. Make sure you check the materialized output to validate that the prompt is getting assembled how you like.</p>
</blockquote>
<p><a id="markdown-2-use-type-inputoutput" name="2-use-type-inputoutput"></a></p>
</section>
<section id="use-type-input_output" class="level3">
<h3 class="anchored" data-anchor-id="use-type-input_output">2. Use <code>type: input_output</code></h3>
<p>Lets materialize data with our <code>output.jsonl</code> file by setting <code>type: input_output</code> in our axolotl config:</p>
<div class="sourceCode" id="cb5"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="co"># training_config.yaml</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="fu">base_model</span><span class="kw">:</span><span class="at"> mistralai/Mistral-7B-v0.1</span></span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a><span class="fu">data_seed</span><span class="kw">:</span><span class="at"> </span><span class="dv">49</span></span>
<span id="cb5-4"><a href="#cb5-4" aria-hidden="true" tabindex="-1"></a><span class="fu">seed</span><span class="kw">:</span><span class="at"> </span><span class="dv">49</span></span>
<span id="cb5-5"><a href="#cb5-5" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-6"><a href="#cb5-6" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb5-7"><a href="#cb5-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> output.jsonl</span></span>
<span id="cb5-8"><a href="#cb5-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> input_output</span></span>
<span id="cb5-9"><a href="#cb5-9" aria-hidden="true" tabindex="-1"></a><span class="fu">val_set_size</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.1</span></span>
<span id="cb5-10"><a href="#cb5-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-11"><a href="#cb5-11" aria-hidden="true" tabindex="-1"></a><span class="fu">sequence_len</span><span class="kw">:</span><span class="at"> </span><span class="dv">896</span></span>
<span id="cb5-12"><a href="#cb5-12" aria-hidden="true" tabindex="-1"></a><span class="fu">sample_packing</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb5-13"><a href="#cb5-13" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-14"><a href="#cb5-14" aria-hidden="true" tabindex="-1"></a><span class="fu">micro_batch_size</span><span class="kw">:</span><span class="at"> </span><span class="dv">2</span></span>
<span id="cb5-15"><a href="#cb5-15" aria-hidden="true" tabindex="-1"></a><span class="fu">gradient_accumulation_steps</span><span class="kw">:</span><span class="at"> </span><span class="dv">3</span></span>
<span id="cb5-16"><a href="#cb5-16" aria-hidden="true" tabindex="-1"></a><span class="fu">eval_batch_size</span><span class="kw">:</span><span class="at"> </span><span class="dv">2</span></span>
<span id="cb5-17"><a href="#cb5-17" aria-hidden="true" tabindex="-1"></a><span class="fu">num_epochs</span><span class="kw">:</span><span class="at"> </span><span class="dv">1</span></span>
<span id="cb5-18"><a href="#cb5-18" aria-hidden="true" tabindex="-1"></a><span class="fu">learning_rate</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.0002</span></span>
<span id="cb5-19"><a href="#cb5-19" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-20"><a href="#cb5-20" aria-hidden="true" tabindex="-1"></a><span class="fu">train_on_inputs</span><span class="kw">:</span><span class="at"> </span><span class="ch">false</span></span>
<span id="cb5-21"><a href="#cb5-21" aria-hidden="true" tabindex="-1"></a><span class="fu">special_tokens</span><span class="kw">:</span></span>
<span id="cb5-22"><a href="#cb5-22" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">bos_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;s&gt;"</span></span>
<span id="cb5-23"><a href="#cb5-23" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">eos_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;/s&gt;"</span></span>
<span id="cb5-24"><a href="#cb5-24" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">unk_token</span><span class="kw">:</span><span class="at"> </span><span class="st">"&lt;unk&gt;"</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>You can use the following command to materialize your data. The <code>--debug</code> flag will print the tokens, along with the labels so you can verify that the correct items are being ignored:</p>
<div class="sourceCode" id="cb6"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="ex">$</span> python <span class="at">-m</span> axolotl.cli.preprocess training_config.yaml <span class="at">--debug</span></span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a><span class="ex">...</span></span>
<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a><span class="ex">[2024-03-05</span> 23:36:46,969] <span class="pp">[</span><span class="ss">INFO</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">axolotl.check_example_labels:35</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">PID:607731</span><span class="pp">]</span> <span class="pp">[</span><span class="ss">RANK:0</span><span class="pp">]</span> <span class="op">&lt;</span>s<span class="op">&gt;(</span><span class="ex">1,</span> 1<span class="op">)</span> Hello<span class="er">(</span><span class="ex">22557,</span> 22557<span class="kw">)</span></span>
<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a><span class="kw">(</span><span class="ex">13,</span> 13<span class="kw">)</span> <span class="ex">hi</span><span class="er">(</span><span class="ex">12014,</span> 12014<span class="kw">)</span> <span class="ex">there</span><span class="er">(</span><span class="ex">736,</span> 736<span class="kw">)</span> <span class="ex">!</span><span class="er">(</span><span class="ex">28808,</span> 28808<span class="kw">)</span> <span class="bu">.</span><span class="er">(</span><span class="ex">28723,</span> 28723<span class="kw">)</span> <span class="kw">(</span><span class="ex">28705,</span> 28705<span class="kw">)</span> <span class="ex">good</span><span class="er">(</span><span class="ex">-100,</span> 1179<span class="kw">)</span> <span class="ex">bye</span><span class="er">(</span><span class="ex">-100,</span> 17664<span class="kw">)</span> <span class="kw">(</span><span class="ex">-100,</span> 28705<span class="kw">)</span> <span class="ex">fare</span><span class="er">(</span><span class="ex">19111,</span> 19111<span class="kw">)</span> <span class="ex">well</span><span class="er">(</span><span class="ex">5458,</span> 5458<span class="kw">)</span> <span class="op">&lt;</span>/s<span class="op">&gt;(</span><span class="ex">2,</span> 2<span class="op">)</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>The format is <code>decoded_token</code>(<code>label</code>, <code>token_id</code>), for example, <code>&lt;s&gt;(1, 1)</code> means that the token is <code>&lt;s&gt;</code>, the label is <code>1</code> and the token_id is <code>1</code>. When the label is <code>-100</code> then that token is ignored for training.</p>
<p><a id="markdown-3-check-the-prompts" name="3-check-the-prompts"></a></p>
</section>
<section id="check-the-prompts" class="level3">
<h3 class="anchored" data-anchor-id="check-the-prompts">3. Check the prompts</h3>
<p>Here is another way to check the materialized output:</p>
<div class="sourceCode" id="cb7"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> transformers <span class="im">import</span> AutoTokenizer</span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> datasets <span class="im">import</span> load_from_disk</span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> yaml</span>
<span id="cb7-4"><a href="#cb7-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-5"><a href="#cb7-5" aria-hidden="true" tabindex="-1"></a>directory <span class="op">=</span> <span class="op">!</span>ls last_run_prepared<span class="op">/</span></span>
<span id="cb7-6"><a href="#cb7-6" aria-hidden="true" tabindex="-1"></a><span class="cf">with</span> <span class="bu">open</span>(<span class="st">'training_config.yaml'</span>, <span class="st">'r'</span>) <span class="im">as</span> f:</span>
<span id="cb7-7"><a href="#cb7-7" aria-hidden="true" tabindex="-1"></a> cfg <span class="op">=</span> yaml.safe_load(f)</span>
<span id="cb7-8"><a href="#cb7-8" aria-hidden="true" tabindex="-1"></a>model_id <span class="op">=</span> cfg[<span class="st">'base_model'</span>]</span>
<span id="cb7-9"><a href="#cb7-9" aria-hidden="true" tabindex="-1"></a>tok <span class="op">=</span> AutoTokenizer.from_pretrained(model_id)</span>
<span id="cb7-10"><a href="#cb7-10" aria-hidden="true" tabindex="-1"></a>ds <span class="op">=</span> load_from_disk(<span class="ss">f'last_run_prepared/</span><span class="sc">{</span>directory[<span class="dv">0</span>]<span class="sc">}</span><span class="ss">/'</span>)</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb8"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="op">&gt;&gt;&gt;</span> row <span class="op">=</span> ds[<span class="dv">0</span>]</span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a><span class="op">&gt;&gt;&gt;</span> <span class="bu">print</span>(tok.decode(row[<span class="st">'input_ids'</span>]))</span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a><span class="op">&lt;</span>s<span class="op">&gt;</span> Hello</span>
<span id="cb8-4"><a href="#cb8-4" aria-hidden="true" tabindex="-1"></a> hi there<span class="op">!</span>. goodbye farewell<span class="op">&lt;/</span>s<span class="op">&gt;</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>We can check that the right tokens are ignored by comparing the labels to each token:</p>
<div class="sourceCode" id="cb9"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> pandas <span class="im">as</span> pd</span>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a>pd.DataFrame([{<span class="st">'token'</span>: tok.decode(i), <span class="st">'label'</span>: l, <span class="st">'id'</span>:i} <span class="cf">for</span> i,l <span class="kw">in</span></span>
<span id="cb9-3"><a href="#cb9-3" aria-hidden="true" tabindex="-1"></a> <span class="bu">zip</span>(row[<span class="st">'input_ids'</span>], row[<span class="st">'labels'</span>])])</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<table class="caption-top table">
<thead>
<tr class="header">
<th>token</th>
<th>label</th>
<th>id</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td>0</td>
<td>&lt;s&gt;</td>
<td>1</td>
</tr>
<tr class="even">
<td>1</td>
<td>Hello</td>
<td>22557</td>
</tr>
<tr class="odd">
<td>2</td>
<td>\n</td>
<td>13</td>
</tr>
<tr class="even">
<td>3</td>
<td>hi</td>
<td>12014</td>
</tr>
<tr class="odd">
<td>4</td>
<td>there</td>
<td>736</td>
</tr>
<tr class="even">
<td>5</td>
<td>!</td>
<td>28808</td>
</tr>
<tr class="odd">
<td>6</td>
<td>.</td>
<td>28723</td>
</tr>
<tr class="even">
<td>7</td>
<td></td>
<td>28705</td>
</tr>
<tr class="odd">
<td>8</td>
<td>good</td>
<td>-100</td>
</tr>
<tr class="even">
<td>9</td>
<td>bye</td>
<td>-100</td>
</tr>
<tr class="odd">
<td>10</td>
<td></td>
<td>-100</td>
</tr>
<tr class="even">
<td>11</td>
<td>fare</td>
<td>19111</td>
</tr>
<tr class="odd">
<td>12</td>
<td>well</td>
<td>5458</td>
</tr>
<tr class="even">
<td>13</td>
<td>&lt;/s&gt;</td>
<td>2</td>
</tr>
</tbody>
</table>
<p>If we look at the input data, the above table seems correct! (The jsonl version is repeated below for reference):</p>
<div class="sourceCode" id="cb10"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="ex">$</span> head <span class="at">-n1</span> output.jsonl <span class="kw">|</span> <span class="ex">python</span> <span class="at">-m</span> json.tool</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="cell-output cell-output-stdout">
<pre><code>{
"segments": [
{
"label": true,
"text": "&lt;s&gt;Hello\n"
},
{
"label": true,
"text": "hi there!. "
},
{
"label": false,
"text": "goodbye "
},
{
"label": true,
"text": "farewell&lt;/s&gt;"
}
]
}</code></pre>
</div>
<p>The documentation moved to <a href="../docs/dataset-formats/template_free.html">here</a>.</p>
</section>
</section>
</main> <!-- /main -->
<script id="quarto-html-after-body" type="application/javascript">

View File

@@ -7,7 +7,7 @@
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<title>Installation Guide Axolotl</title>
<title>Installation Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/installation.html">Installation Guide</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/installation.html">Installation</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,18 +240,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -356,9 +459,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/installation.html">Installation Guide</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">Getting Started</a></li><li class="breadcrumb-item"><a href="../docs/installation.html">Installation</a></li></ol></nav>
<div class="quarto-title">
<div class="quarto-title-block"><div><h1 class="title">Installation Guide</h1><button type="button" class="btn code-tools-button" id="quarto-code-tools-source"><i class="bi"></i> Code</button></div></div>
<h1 class="title">Installation</h1>
</div>
@@ -488,8 +591,6 @@ Important
<p>If you encounter installation issues, see our <a href="../docs/faq.html">FAQ</a> and <a href="../docs/debugging.html">Debugging Guide</a>.</p>
<!-- -->
</section>
</main> <!-- /main -->
@@ -580,57 +681,6 @@ window.document.addEventListener("DOMContentLoaded", function (event) {
container: window.document.getElementById('quarto-embedded-source-code-modal')
});
clipboardModal.on('success', onCopySuccess);
}
const viewSource = window.document.getElementById('quarto-view-source') ||
window.document.getElementById('quarto-code-tools-source');
if (viewSource) {
const sourceUrl = viewSource.getAttribute("data-quarto-source-url");
viewSource.addEventListener("click", function(e) {
if (sourceUrl) {
// rstudio viewer pane
if (/\bcapabilities=\b/.test(window.location)) {
window.open(sourceUrl);
} else {
window.location.href = sourceUrl;
}
} else {
const modal = new bootstrap.Modal(document.getElementById('quarto-embedded-source-code-modal'));
modal.show();
}
return false;
});
}
function toggleCodeHandler(show) {
return function(e) {
const detailsSrc = window.document.querySelectorAll(".cell > details > .sourceCode");
for (let i=0; i<detailsSrc.length; i++) {
const details = detailsSrc[i].parentElement;
if (show) {
details.open = true;
} else {
details.removeAttribute("open");
}
}
const cellCodeDivs = window.document.querySelectorAll(".cell > .sourceCode");
const fromCls = show ? "hidden" : "unhidden";
const toCls = show ? "unhidden" : "hidden";
for (let i=0; i<cellCodeDivs.length; i++) {
const codeDiv = cellCodeDivs[i];
if (codeDiv.classList.contains(fromCls)) {
codeDiv.classList.remove(fromCls);
codeDiv.classList.add(toCls);
}
}
return false;
}
}
const hideAllCode = window.document.getElementById("quarto-hide-all-code");
if (hideAllCode) {
hideAllCode.addEventListener("click", toggleCodeHandler(false));
}
const showAllCode = window.document.getElementById("quarto-show-all-code");
if (showAllCode) {
showAllCode.addEventListener("click", toggleCodeHandler(true));
}
var localhostRegex = new RegExp(/^(?:http|https):\/\/localhost\:?[0-9]*\//);
var mailtoRegex = new RegExp(/^mailto:/);
@@ -960,127 +1010,7 @@ window.document.addEventListener("DOMContentLoaded", function (event) {
}
}
});
</script><div class="modal fade" id="quarto-embedded-source-code-modal" tabindex="-1" aria-labelledby="quarto-embedded-source-code-modal-label" aria-hidden="true"><div class="modal-dialog modal-dialog-scrollable"><div class="modal-content"><div class="modal-header"><h5 class="modal-title" id="quarto-embedded-source-code-modal-label">Source Code</h5><button class="btn-close" data-bs-dismiss="modal"></button></div><div class="modal-body"><div class="">
<div class="sourceCode" id="cb9" data-shortcodes="false"><pre class="sourceCode markdown code-with-copy"><code class="sourceCode markdown"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a><span class="an">title:</span><span class="co"> "Installation Guide"</span></span>
<span id="cb9-3"><a href="#cb9-3" aria-hidden="true" tabindex="-1"></a><span class="an">format:</span></span>
<span id="cb9-4"><a href="#cb9-4" aria-hidden="true" tabindex="-1"></a><span class="co"> html:</span></span>
<span id="cb9-5"><a href="#cb9-5" aria-hidden="true" tabindex="-1"></a><span class="co"> toc: true</span></span>
<span id="cb9-6"><a href="#cb9-6" aria-hidden="true" tabindex="-1"></a><span class="co"> toc-depth: 3</span></span>
<span id="cb9-7"><a href="#cb9-7" aria-hidden="true" tabindex="-1"></a><span class="co"> number-sections: true</span></span>
<span id="cb9-8"><a href="#cb9-8" aria-hidden="true" tabindex="-1"></a><span class="co"> code-tools: true</span></span>
<span id="cb9-9"><a href="#cb9-9" aria-hidden="true" tabindex="-1"></a><span class="an">execute:</span></span>
<span id="cb9-10"><a href="#cb9-10" aria-hidden="true" tabindex="-1"></a><span class="co"> enabled: false</span></span>
<span id="cb9-11"><a href="#cb9-11" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb9-12"><a href="#cb9-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-13"><a href="#cb9-13" aria-hidden="true" tabindex="-1"></a>This guide covers all the ways you can install and set up Axolotl for your environment.</span>
<span id="cb9-14"><a href="#cb9-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-15"><a href="#cb9-15" aria-hidden="true" tabindex="-1"></a><span class="fu">## Requirements {#sec-requirements}</span></span>
<span id="cb9-16"><a href="#cb9-16" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-17"><a href="#cb9-17" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>NVIDIA GPU (Ampere architecture or newer for <span class="in">`bf16`</span> and Flash Attention) or AMD GPU</span>
<span id="cb9-18"><a href="#cb9-18" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Python ≥3.10</span>
<span id="cb9-19"><a href="#cb9-19" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>PyTorch ≥2.4.1</span>
<span id="cb9-20"><a href="#cb9-20" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-21"><a href="#cb9-21" aria-hidden="true" tabindex="-1"></a><span class="fu">## Installation Methods {#sec-installation-methods}</span></span>
<span id="cb9-22"><a href="#cb9-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-23"><a href="#cb9-23" aria-hidden="true" tabindex="-1"></a><span class="fu">### PyPI Installation (Recommended) {#sec-pypi}</span></span>
<span id="cb9-24"><a href="#cb9-24" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-25"><a href="#cb9-25" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-26"><a href="#cb9-26" aria-hidden="true" tabindex="-1"></a><span class="in">pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]</span></span>
<span id="cb9-27"><a href="#cb9-27" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-28"><a href="#cb9-28" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-29"><a href="#cb9-29" aria-hidden="true" tabindex="-1"></a>We use <span class="in">`--no-build-isolation`</span> in order to detect the installed PyTorch version (if</span>
<span id="cb9-30"><a href="#cb9-30" aria-hidden="true" tabindex="-1"></a>installed) in order not to clobber it, and so that we set the correct version of</span>
<span id="cb9-31"><a href="#cb9-31" aria-hidden="true" tabindex="-1"></a>dependencies that are specific to the PyTorch version or other installed</span>
<span id="cb9-32"><a href="#cb9-32" aria-hidden="true" tabindex="-1"></a>co-dependencies.</span>
<span id="cb9-33"><a href="#cb9-33" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-34"><a href="#cb9-34" aria-hidden="true" tabindex="-1"></a><span class="fu">### Edge/Development Build {#sec-edge-build}</span></span>
<span id="cb9-35"><a href="#cb9-35" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-36"><a href="#cb9-36" aria-hidden="true" tabindex="-1"></a>For the latest features between releases:</span>
<span id="cb9-37"><a href="#cb9-37" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-38"><a href="#cb9-38" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-39"><a href="#cb9-39" aria-hidden="true" tabindex="-1"></a><span class="in">git clone https://github.com/axolotl-ai-cloud/axolotl.git</span></span>
<span id="cb9-40"><a href="#cb9-40" aria-hidden="true" tabindex="-1"></a><span class="in">cd axolotl</span></span>
<span id="cb9-41"><a href="#cb9-41" aria-hidden="true" tabindex="-1"></a><span class="in">pip3 install packaging ninja</span></span>
<span id="cb9-42"><a href="#cb9-42" aria-hidden="true" tabindex="-1"></a><span class="in">pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'</span></span>
<span id="cb9-43"><a href="#cb9-43" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-44"><a href="#cb9-44" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-45"><a href="#cb9-45" aria-hidden="true" tabindex="-1"></a><span class="fu">### Docker {#sec-docker}</span></span>
<span id="cb9-46"><a href="#cb9-46" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-47"><a href="#cb9-47" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-48"><a href="#cb9-48" aria-hidden="true" tabindex="-1"></a><span class="in">docker run --gpus '"all"' --rm -it axolotlai/axolotl:main-latest</span></span>
<span id="cb9-49"><a href="#cb9-49" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-50"><a href="#cb9-50" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-51"><a href="#cb9-51" aria-hidden="true" tabindex="-1"></a>For development with Docker:</span>
<span id="cb9-52"><a href="#cb9-52" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-53"><a href="#cb9-53" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-54"><a href="#cb9-54" aria-hidden="true" tabindex="-1"></a><span class="in">docker compose up -d</span></span>
<span id="cb9-55"><a href="#cb9-55" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-56"><a href="#cb9-56" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-57"><a href="#cb9-57" aria-hidden="true" tabindex="-1"></a>::: {.callout-tip}</span>
<span id="cb9-58"><a href="#cb9-58" aria-hidden="true" tabindex="-1"></a><span class="fu">### Advanced Docker Configuration</span></span>
<span id="cb9-59"><a href="#cb9-59" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-60"><a href="#cb9-60" aria-hidden="true" tabindex="-1"></a><span class="in">docker run --privileged --gpus '"all"' --shm-size 10g --rm -it \</span></span>
<span id="cb9-61"><a href="#cb9-61" aria-hidden="true" tabindex="-1"></a><span class="in"> --name axolotl --ipc=host \</span></span>
<span id="cb9-62"><a href="#cb9-62" aria-hidden="true" tabindex="-1"></a><span class="in"> --ulimit memlock=-1 --ulimit stack=67108864 \</span></span>
<span id="cb9-63"><a href="#cb9-63" aria-hidden="true" tabindex="-1"></a><span class="in"> --mount type=bind,src="${PWD}",target=/workspace/axolotl \</span></span>
<span id="cb9-64"><a href="#cb9-64" aria-hidden="true" tabindex="-1"></a><span class="in"> -v ${HOME}/.cache/huggingface:/root/.cache/huggingface \</span></span>
<span id="cb9-65"><a href="#cb9-65" aria-hidden="true" tabindex="-1"></a><span class="in"> axolotlai/axolotl:main-latest</span></span>
<span id="cb9-66"><a href="#cb9-66" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-67"><a href="#cb9-67" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb9-68"><a href="#cb9-68" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-69"><a href="#cb9-69" aria-hidden="true" tabindex="-1"></a><span class="fu">## Cloud Environments {#sec-cloud}</span></span>
<span id="cb9-70"><a href="#cb9-70" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-71"><a href="#cb9-71" aria-hidden="true" tabindex="-1"></a><span class="fu">### Cloud GPU Providers {#sec-cloud-gpu}</span></span>
<span id="cb9-72"><a href="#cb9-72" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-73"><a href="#cb9-73" aria-hidden="true" tabindex="-1"></a>For providers supporting Docker:</span>
<span id="cb9-74"><a href="#cb9-74" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-75"><a href="#cb9-75" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Use <span class="in">`axolotlai/axolotl-cloud:main-latest`</span></span>
<span id="cb9-76"><a href="#cb9-76" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Available on:</span>
<span id="cb9-77"><a href="#cb9-77" aria-hidden="true" tabindex="-1"></a><span class="ss"> - </span><span class="co">[</span><span class="ot">Latitude.sh</span><span class="co">](https://latitude.sh/blueprint/989e0e79-3bf6-41ea-a46b-1f246e309d5c)</span></span>
<span id="cb9-78"><a href="#cb9-78" aria-hidden="true" tabindex="-1"></a><span class="ss"> - </span><span class="co">[</span><span class="ot">JarvisLabs.ai</span><span class="co">](https://jarvislabs.ai/templates/axolotl)</span></span>
<span id="cb9-79"><a href="#cb9-79" aria-hidden="true" tabindex="-1"></a><span class="ss"> - </span><span class="co">[</span><span class="ot">RunPod</span><span class="co">](https://runpod.io/gsc?template=v2ickqhz9s&amp;ref=6i7fkpdz)</span></span>
<span id="cb9-80"><a href="#cb9-80" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-81"><a href="#cb9-81" aria-hidden="true" tabindex="-1"></a><span class="fu">### Google Colab {#sec-colab}</span></span>
<span id="cb9-82"><a href="#cb9-82" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-83"><a href="#cb9-83" aria-hidden="true" tabindex="-1"></a>Use our <span class="co">[</span><span class="ot">example notebook</span><span class="co">](../examples/colab-notebooks/colab-axolotl-example.ipynb)</span>.</span>
<span id="cb9-84"><a href="#cb9-84" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-85"><a href="#cb9-85" aria-hidden="true" tabindex="-1"></a><span class="fu">## Platform-Specific Instructions {#sec-platform-specific}</span></span>
<span id="cb9-86"><a href="#cb9-86" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-87"><a href="#cb9-87" aria-hidden="true" tabindex="-1"></a><span class="fu">### macOS {#sec-macos}</span></span>
<span id="cb9-88"><a href="#cb9-88" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-89"><a href="#cb9-89" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb9-90"><a href="#cb9-90" aria-hidden="true" tabindex="-1"></a><span class="in">pip3 install --no-build-isolation -e '.'</span></span>
<span id="cb9-91"><a href="#cb9-91" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb9-92"><a href="#cb9-92" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-93"><a href="#cb9-93" aria-hidden="true" tabindex="-1"></a>See @sec-troubleshooting for Mac-specific issues.</span>
<span id="cb9-94"><a href="#cb9-94" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-95"><a href="#cb9-95" aria-hidden="true" tabindex="-1"></a><span class="fu">### Windows {#sec-windows}</span></span>
<span id="cb9-96"><a href="#cb9-96" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-97"><a href="#cb9-97" aria-hidden="true" tabindex="-1"></a>::: {.callout-important}</span>
<span id="cb9-98"><a href="#cb9-98" aria-hidden="true" tabindex="-1"></a>We recommend using WSL2 (Windows Subsystem for Linux) or Docker.</span>
<span id="cb9-99"><a href="#cb9-99" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb9-100"><a href="#cb9-100" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-101"><a href="#cb9-101" aria-hidden="true" tabindex="-1"></a><span class="fu">## Environment Managers {#sec-env-managers}</span></span>
<span id="cb9-102"><a href="#cb9-102" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-103"><a href="#cb9-103" aria-hidden="true" tabindex="-1"></a><span class="fu">### Conda/Pip venv {#sec-conda}</span></span>
<span id="cb9-104"><a href="#cb9-104" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-105"><a href="#cb9-105" aria-hidden="true" tabindex="-1"></a><span class="ss">1. </span>Install Python ≥3.10</span>
<span id="cb9-106"><a href="#cb9-106" aria-hidden="true" tabindex="-1"></a><span class="ss">2. </span>Install PyTorch: https://pytorch.org/get-started/locally/</span>
<span id="cb9-107"><a href="#cb9-107" aria-hidden="true" tabindex="-1"></a><span class="ss">3. </span>Install Axolotl:</span>
<span id="cb9-108"><a href="#cb9-108" aria-hidden="true" tabindex="-1"></a> <span class="in">```{.bash}</span></span>
<span id="cb9-109"><a href="#cb9-109" aria-hidden="true" tabindex="-1"></a><span class="in"> pip3 install packaging</span></span>
<span id="cb9-110"><a href="#cb9-110" aria-hidden="true" tabindex="-1"></a><span class="in"> pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'</span></span>
<span id="cb9-111"><a href="#cb9-111" aria-hidden="true" tabindex="-1"></a><span class="in"> ```</span></span>
<span id="cb9-112"><a href="#cb9-112" aria-hidden="true" tabindex="-1"></a><span class="ss">4. </span>(Optional) Login to Hugging Face:</span>
<span id="cb9-113"><a href="#cb9-113" aria-hidden="true" tabindex="-1"></a> <span class="in">```{.bash}</span></span>
<span id="cb9-114"><a href="#cb9-114" aria-hidden="true" tabindex="-1"></a><span class="in"> huggingface-cli login</span></span>
<span id="cb9-115"><a href="#cb9-115" aria-hidden="true" tabindex="-1"></a><span class="in"> ```</span></span>
<span id="cb9-116"><a href="#cb9-116" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-117"><a href="#cb9-117" aria-hidden="true" tabindex="-1"></a><span class="fu">## Troubleshooting {#sec-troubleshooting}</span></span>
<span id="cb9-118"><a href="#cb9-118" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb9-119"><a href="#cb9-119" aria-hidden="true" tabindex="-1"></a>If you encounter installation issues, see our <span class="co">[</span><span class="ot">FAQ</span><span class="co">](faq.qmd)</span> and <span class="co">[</span><span class="ot">Debugging Guide</span><span class="co">](debugging.qmd)</span>.</span></code><button title="Copy to Clipboard" class="code-copy-button" data-in-quarto-modal=""><i class="bi"></i></button></pre></div>
</div></div></div></div></div>
</script>
</div> <!-- /content -->

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">LoRA Optimizations</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/lora_optims.html">LoRA Optimizations</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -344,7 +447,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/lora_optims.html">LoRA Optimizations</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">LoRA Optimizations</h1>
</div>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">Learning Rate Groups</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/lr_groups.html">Learning Rate Groups</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -337,7 +440,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/lr_groups.html">Learning Rate Groups</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Learning Rate Groups</h1>
</div>

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/mac.html">Mac M-series</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/mac.html">Mac M-series</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +207,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -296,7 +399,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/mac.html">Mac M-series</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/mac.html">Mac M-series</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Mac M-series</h1>
</div>
@@ -334,7 +437,10 @@ ul.task-list li input[type="checkbox"] {
<li><label><input type="checkbox">qlora</label></li>
<li><label><input type="checkbox">DeepSpeed</label></li>
</ul>
<p>Untested: - FSDP</p>
<p>Untested:</p>
<ul>
<li>FSDP</li>
</ul>

View File

@@ -7,7 +7,7 @@
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
<title>Multi-GPU Training Guide Axolotl</title>
<title>Multi-GPU Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Multi-GPU Training Guide</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Multi-GPU</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,18 +240,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -355,9 +458,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Multi-GPU Training Guide</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Multi-GPU</a></li></ol></nav>
<div class="quarto-title">
<div class="quarto-title-block"><div><h1 class="title">Multi-GPU Training Guide</h1><button type="button" class="btn code-tools-button" id="quarto-code-tools-source"><i class="bi"></i> Code</button></div></div>
<div class="quarto-title-block"><div><h1 class="title">Multi-GPU</h1><button type="button" class="btn code-tools-button" id="quarto-code-tools-source"><i class="bi"></i> Code</button></div></div>
</div>
@@ -394,7 +497,11 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
</section>
<section id="sec-deepspeed-usage" class="level3" data-number="2.2">
<h3 data-number="2.2" class="anchored" data-anchor-id="sec-deepspeed-usage"><span class="header-section-number">2.2</span> Usage</h3>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="ex">accelerate</span> launch <span class="at">-m</span> axolotl.cli.train examples/llama-2/config.yml <span class="at">--deepspeed</span> deepspeed_configs/zero1.json</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Passing arg via config</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml</span>
<span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb2-4"><a href="#cb2-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Passing arg via cli</span></span>
<span id="cb2-5"><a href="#cb2-5" aria-hidden="true" tabindex="-1"></a><span class="ex">axolotl</span> train config.yml <span class="at">--deepspeed</span> deepspeed_configs/zero1.json</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="sec-zero-stages" class="level3" data-number="2.3">
<h3 data-number="2.3" class="anchored" data-anchor-id="sec-zero-stages"><span class="header-section-number">2.3</span> ZeRO Stages</h3>
@@ -428,32 +535,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<h2 data-number="4" class="anchored" data-anchor-id="sec-performance"><span class="header-section-number">4</span> Performance Optimization</h2>
<section id="sec-liger" class="level3" data-number="4.1">
<h3 data-number="4.1" class="anchored" data-anchor-id="sec-liger"><span class="header-section-number">4.1</span> Liger Kernel Integration</h3>
<div class="callout callout-style-default callout-note callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Note
</div>
</div>
<div class="callout-body-container callout-body">
<p>Liger Kernel provides efficient Triton kernels for LLM training, offering:</p>
<ul>
<li>20% increase in multi-GPU training throughput</li>
<li>60% reduction in memory usage</li>
<li>Compatibility with both FSDP and DeepSpeed</li>
</ul>
</div>
</div>
<p>Configuration:</p>
<div class="sourceCode" id="cb4"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="fu">plugins</span><span class="kw">:</span></span>
<span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> axolotl.integrations.liger.LigerPlugin</span></span>
<span id="cb4-3"><a href="#cb4-3" aria-hidden="true" tabindex="-1"></a><span class="fu">liger_rope</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb4-4"><a href="#cb4-4" aria-hidden="true" tabindex="-1"></a><span class="fu">liger_rms_norm</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb4-5"><a href="#cb4-5" aria-hidden="true" tabindex="-1"></a><span class="fu">liger_glu_activation</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb4-6"><a href="#cb4-6" aria-hidden="true" tabindex="-1"></a><span class="fu">liger_layer_norm</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb4-7"><a href="#cb4-7" aria-hidden="true" tabindex="-1"></a><span class="fu">liger_fused_linear_cross_entropy</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Please see <a href="../docs/custom_integrations.html#liger">docs</a> for more info.</p>
</section>
</section>
<section id="sec-troubleshooting" class="level2" data-number="5">
@@ -961,124 +1043,110 @@ window.document.addEventListener("DOMContentLoaded", function (event) {
}
});
</script><div class="modal fade" id="quarto-embedded-source-code-modal" tabindex="-1" aria-labelledby="quarto-embedded-source-code-modal-label" aria-hidden="true"><div class="modal-dialog modal-dialog-scrollable"><div class="modal-content"><div class="modal-header"><h5 class="modal-title" id="quarto-embedded-source-code-modal-label">Source Code</h5><button class="btn-close" data-bs-dismiss="modal"></button></div><div class="modal-body"><div class="">
<div class="sourceCode" id="cb5" data-shortcodes="false"><pre class="sourceCode markdown code-with-copy"><code class="sourceCode markdown"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="an">title:</span><span class="co"> "Multi-GPU Training Guide"</span></span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a><span class="an">format:</span></span>
<span id="cb5-4"><a href="#cb5-4" aria-hidden="true" tabindex="-1"></a><span class="co"> html:</span></span>
<span id="cb5-5"><a href="#cb5-5" aria-hidden="true" tabindex="-1"></a><span class="co"> toc: true</span></span>
<span id="cb5-6"><a href="#cb5-6" aria-hidden="true" tabindex="-1"></a><span class="co"> toc-depth: 3</span></span>
<span id="cb5-7"><a href="#cb5-7" aria-hidden="true" tabindex="-1"></a><span class="co"> number-sections: true</span></span>
<span id="cb5-8"><a href="#cb5-8" aria-hidden="true" tabindex="-1"></a><span class="co"> code-tools: true</span></span>
<span id="cb5-9"><a href="#cb5-9" aria-hidden="true" tabindex="-1"></a><span class="an">execute:</span></span>
<span id="cb5-10"><a href="#cb5-10" aria-hidden="true" tabindex="-1"></a><span class="co"> enabled: false</span></span>
<span id="cb5-11"><a href="#cb5-11" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb5-12"><a href="#cb5-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-13"><a href="#cb5-13" aria-hidden="true" tabindex="-1"></a>This guide covers advanced training configurations for multi-GPU setups using Axolotl.</span>
<span id="cb5-14"><a href="#cb5-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-15"><a href="#cb5-15" aria-hidden="true" tabindex="-1"></a><span class="fu">## Overview {#sec-overview}</span></span>
<span id="cb5-16"><a href="#cb5-16" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-17"><a href="#cb5-17" aria-hidden="true" tabindex="-1"></a>Axolotl supports several methods for multi-GPU training:</span>
<span id="cb5-18"><a href="#cb5-18" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-19"><a href="#cb5-19" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>DeepSpeed (recommended)</span>
<span id="cb5-20"><a href="#cb5-20" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>FSDP (Fully Sharded Data Parallel)</span>
<span id="cb5-21"><a href="#cb5-21" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>FSDP + QLoRA</span>
<span id="cb5-22"><a href="#cb5-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-23"><a href="#cb5-23" aria-hidden="true" tabindex="-1"></a><span class="fu">## DeepSpeed {#sec-deepspeed}</span></span>
<span id="cb5-24"><a href="#cb5-24" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-25"><a href="#cb5-25" aria-hidden="true" tabindex="-1"></a>DeepSpeed is the recommended approach for multi-GPU training due to its stability and performance. It provides various optimization levels through ZeRO stages.</span>
<span id="cb5-26"><a href="#cb5-26" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-27"><a href="#cb5-27" aria-hidden="true" tabindex="-1"></a><span class="fu">### Configuration {#sec-deepspeed-config}</span></span>
<span id="cb5-28"><a href="#cb5-28" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-29"><a href="#cb5-29" aria-hidden="true" tabindex="-1"></a>Add to your YAML config:</span>
<span id="cb5-30"><a href="#cb5-30" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-31"><a href="#cb5-31" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb5-32"><a href="#cb5-32" aria-hidden="true" tabindex="-1"></a><span class="in">deepspeed: deepspeed_configs/zero1.json</span></span>
<span id="cb5-33"><a href="#cb5-33" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb5-34"><a href="#cb5-34" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-35"><a href="#cb5-35" aria-hidden="true" tabindex="-1"></a><span class="fu">### Usage {#sec-deepspeed-usage}</span></span>
<span id="cb5-36"><a href="#cb5-36" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-37"><a href="#cb5-37" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb5-38"><a href="#cb5-38" aria-hidden="true" tabindex="-1"></a><span class="in">accelerate launch -m axolotl.cli.train examples/llama-2/config.yml --deepspeed deepspeed_configs/zero1.json</span></span>
<span id="cb5-39"><a href="#cb5-39" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb5-40"><a href="#cb5-40" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-41"><a href="#cb5-41" aria-hidden="true" tabindex="-1"></a><span class="fu">### ZeRO Stages {#sec-zero-stages}</span></span>
<span id="cb5-42"><a href="#cb5-42" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-43"><a href="#cb5-43" aria-hidden="true" tabindex="-1"></a>We provide default configurations for:</span>
<span id="cb5-44"><a href="#cb5-44" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-45"><a href="#cb5-45" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 1 (<span class="in">`zero1.json`</span>)</span>
<span id="cb5-46"><a href="#cb5-46" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 2 (<span class="in">`zero2.json`</span>)</span>
<span id="cb5-47"><a href="#cb5-47" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 3 (<span class="in">`zero3.json`</span>)</span>
<span id="cb5-48"><a href="#cb5-48" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-49"><a href="#cb5-49" aria-hidden="true" tabindex="-1"></a>Choose based on your memory requirements and performance needs.</span>
<span id="cb5-50"><a href="#cb5-50" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-51"><a href="#cb5-51" aria-hidden="true" tabindex="-1"></a><span class="fu">## FSDP {#sec-fsdp}</span></span>
<span id="cb5-52"><a href="#cb5-52" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-53"><a href="#cb5-53" aria-hidden="true" tabindex="-1"></a><span class="fu">### Basic FSDP Configuration {#sec-fsdp-config}</span></span>
<span id="cb5-54"><a href="#cb5-54" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-55"><a href="#cb5-55" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb5-56"><a href="#cb5-56" aria-hidden="true" tabindex="-1"></a><span class="in">fsdp:</span></span>
<span id="cb5-57"><a href="#cb5-57" aria-hidden="true" tabindex="-1"></a><span class="in"> - full_shard</span></span>
<span id="cb5-58"><a href="#cb5-58" aria-hidden="true" tabindex="-1"></a><span class="in"> - auto_wrap</span></span>
<span id="cb5-59"><a href="#cb5-59" aria-hidden="true" tabindex="-1"></a><span class="in">fsdp_config:</span></span>
<span id="cb5-60"><a href="#cb5-60" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_offload_params: true</span></span>
<span id="cb5-61"><a href="#cb5-61" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_state_dict_type: FULL_STATE_DICT</span></span>
<span id="cb5-62"><a href="#cb5-62" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer</span></span>
<span id="cb5-63"><a href="#cb5-63" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb5-64"><a href="#cb5-64" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-65"><a href="#cb5-65" aria-hidden="true" tabindex="-1"></a><span class="fu">### FSDP + QLoRA {#sec-fsdp-qlora}</span></span>
<span id="cb5-66"><a href="#cb5-66" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-67"><a href="#cb5-67" aria-hidden="true" tabindex="-1"></a>For combining FSDP with QLoRA, see our <span class="co">[</span><span class="ot">dedicated guide</span><span class="co">](fsdp_qlora.qmd)</span>.</span>
<span id="cb5-68"><a href="#cb5-68" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-69"><a href="#cb5-69" aria-hidden="true" tabindex="-1"></a><span class="fu">## Performance Optimization {#sec-performance}</span></span>
<span id="cb5-70"><a href="#cb5-70" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-71"><a href="#cb5-71" aria-hidden="true" tabindex="-1"></a><span class="fu">### Liger Kernel Integration {#sec-liger}</span></span>
<span id="cb5-72"><a href="#cb5-72" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-73"><a href="#cb5-73" aria-hidden="true" tabindex="-1"></a>::: {.callout-note}</span>
<span id="cb5-74"><a href="#cb5-74" aria-hidden="true" tabindex="-1"></a>Liger Kernel provides efficient Triton kernels for LLM training, offering:</span>
<span id="cb5-75"><a href="#cb5-75" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-76"><a href="#cb5-76" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>20% increase in multi-GPU training throughput</span>
<span id="cb5-77"><a href="#cb5-77" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>60% reduction in memory usage</span>
<span id="cb5-78"><a href="#cb5-78" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Compatibility with both FSDP and DeepSpeed</span>
<span id="cb5-79"><a href="#cb5-79" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb5-80"><a href="#cb5-80" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-81"><a href="#cb5-81" aria-hidden="true" tabindex="-1"></a>Configuration:</span>
<span id="cb5-82"><a href="#cb5-82" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-83"><a href="#cb5-83" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb5-84"><a href="#cb5-84" aria-hidden="true" tabindex="-1"></a><span class="in">plugins:</span></span>
<span id="cb5-85"><a href="#cb5-85" aria-hidden="true" tabindex="-1"></a><span class="in"> - axolotl.integrations.liger.LigerPlugin</span></span>
<span id="cb5-86"><a href="#cb5-86" aria-hidden="true" tabindex="-1"></a><span class="in">liger_rope: true</span></span>
<span id="cb5-87"><a href="#cb5-87" aria-hidden="true" tabindex="-1"></a><span class="in">liger_rms_norm: true</span></span>
<span id="cb5-88"><a href="#cb5-88" aria-hidden="true" tabindex="-1"></a><span class="in">liger_glu_activation: true</span></span>
<span id="cb5-89"><a href="#cb5-89" aria-hidden="true" tabindex="-1"></a><span class="in">liger_layer_norm: true</span></span>
<span id="cb5-90"><a href="#cb5-90" aria-hidden="true" tabindex="-1"></a><span class="in">liger_fused_linear_cross_entropy: true</span></span>
<span id="cb5-91"><a href="#cb5-91" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb5-92"><a href="#cb5-92" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-93"><a href="#cb5-93" aria-hidden="true" tabindex="-1"></a><span class="fu">## Troubleshooting {#sec-troubleshooting}</span></span>
<span id="cb5-94"><a href="#cb5-94" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-95"><a href="#cb5-95" aria-hidden="true" tabindex="-1"></a><span class="fu">### NCCL Issues {#sec-nccl}</span></span>
<span id="cb5-96"><a href="#cb5-96" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-97"><a href="#cb5-97" aria-hidden="true" tabindex="-1"></a>For NCCL-related problems, see our <span class="co">[</span><span class="ot">NCCL troubleshooting guide</span><span class="co">](nccl.qmd)</span>.</span>
<span id="cb5-98"><a href="#cb5-98" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-99"><a href="#cb5-99" aria-hidden="true" tabindex="-1"></a><span class="fu">### Common Problems {#sec-common-problems}</span></span>
<span id="cb5-100"><a href="#cb5-100" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-101"><a href="#cb5-101" aria-hidden="true" tabindex="-1"></a>::: {.panel-tabset}</span>
<span id="cb5-102"><a href="#cb5-102" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-103"><a href="#cb5-103" aria-hidden="true" tabindex="-1"></a><span class="fu">## Memory Issues</span></span>
<span id="cb5-104"><a href="#cb5-104" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-105"><a href="#cb5-105" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Reduce <span class="in">`micro_batch_size`</span></span>
<span id="cb5-106"><a href="#cb5-106" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Reduce <span class="in">`eval_batch_size`</span></span>
<span id="cb5-107"><a href="#cb5-107" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Adjust <span class="in">`gradient_accumulation_steps`</span></span>
<span id="cb5-108"><a href="#cb5-108" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Consider using a higher ZeRO stage</span>
<span id="cb5-109"><a href="#cb5-109" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-110"><a href="#cb5-110" aria-hidden="true" tabindex="-1"></a><span class="fu">## Training Instability</span></span>
<span id="cb5-111"><a href="#cb5-111" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-112"><a href="#cb5-112" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Start with DeepSpeed ZeRO-2</span>
<span id="cb5-113"><a href="#cb5-113" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Monitor loss values</span>
<span id="cb5-114"><a href="#cb5-114" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Check learning rates</span>
<span id="cb5-115"><a href="#cb5-115" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-116"><a href="#cb5-116" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb5-117"><a href="#cb5-117" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb5-118"><a href="#cb5-118" aria-hidden="true" tabindex="-1"></a>For more detailed troubleshooting, see our <span class="co">[</span><span class="ot">debugging guide</span><span class="co">](debugging.qmd)</span>.</span></code><button title="Copy to Clipboard" class="code-copy-button" data-in-quarto-modal=""><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb4" data-shortcodes="false"><pre class="sourceCode markdown code-with-copy"><code class="sourceCode markdown"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a><span class="an">title:</span><span class="co"> "Multi-GPU"</span></span>
<span id="cb4-3"><a href="#cb4-3" aria-hidden="true" tabindex="-1"></a><span class="an">format:</span></span>
<span id="cb4-4"><a href="#cb4-4" aria-hidden="true" tabindex="-1"></a><span class="co"> html:</span></span>
<span id="cb4-5"><a href="#cb4-5" aria-hidden="true" tabindex="-1"></a><span class="co"> toc: true</span></span>
<span id="cb4-6"><a href="#cb4-6" aria-hidden="true" tabindex="-1"></a><span class="co"> toc-depth: 3</span></span>
<span id="cb4-7"><a href="#cb4-7" aria-hidden="true" tabindex="-1"></a><span class="co"> number-sections: true</span></span>
<span id="cb4-8"><a href="#cb4-8" aria-hidden="true" tabindex="-1"></a><span class="co"> code-tools: true</span></span>
<span id="cb4-9"><a href="#cb4-9" aria-hidden="true" tabindex="-1"></a><span class="an">execute:</span></span>
<span id="cb4-10"><a href="#cb4-10" aria-hidden="true" tabindex="-1"></a><span class="co"> enabled: false</span></span>
<span id="cb4-11"><a href="#cb4-11" aria-hidden="true" tabindex="-1"></a><span class="co">---</span></span>
<span id="cb4-12"><a href="#cb4-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-13"><a href="#cb4-13" aria-hidden="true" tabindex="-1"></a>This guide covers advanced training configurations for multi-GPU setups using Axolotl.</span>
<span id="cb4-14"><a href="#cb4-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-15"><a href="#cb4-15" aria-hidden="true" tabindex="-1"></a><span class="fu">## Overview {#sec-overview}</span></span>
<span id="cb4-16"><a href="#cb4-16" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-17"><a href="#cb4-17" aria-hidden="true" tabindex="-1"></a>Axolotl supports several methods for multi-GPU training:</span>
<span id="cb4-18"><a href="#cb4-18" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-19"><a href="#cb4-19" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>DeepSpeed (recommended)</span>
<span id="cb4-20"><a href="#cb4-20" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>FSDP (Fully Sharded Data Parallel)</span>
<span id="cb4-21"><a href="#cb4-21" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>FSDP + QLoRA</span>
<span id="cb4-22"><a href="#cb4-22" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-23"><a href="#cb4-23" aria-hidden="true" tabindex="-1"></a><span class="fu">## DeepSpeed {#sec-deepspeed}</span></span>
<span id="cb4-24"><a href="#cb4-24" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-25"><a href="#cb4-25" aria-hidden="true" tabindex="-1"></a>DeepSpeed is the recommended approach for multi-GPU training due to its stability and performance. It provides various optimization levels through ZeRO stages.</span>
<span id="cb4-26"><a href="#cb4-26" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-27"><a href="#cb4-27" aria-hidden="true" tabindex="-1"></a><span class="fu">### Configuration {#sec-deepspeed-config}</span></span>
<span id="cb4-28"><a href="#cb4-28" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-29"><a href="#cb4-29" aria-hidden="true" tabindex="-1"></a>Add to your YAML config:</span>
<span id="cb4-30"><a href="#cb4-30" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-31"><a href="#cb4-31" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb4-32"><a href="#cb4-32" aria-hidden="true" tabindex="-1"></a><span class="in">deepspeed: deepspeed_configs/zero1.json</span></span>
<span id="cb4-33"><a href="#cb4-33" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb4-34"><a href="#cb4-34" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-35"><a href="#cb4-35" aria-hidden="true" tabindex="-1"></a><span class="fu">### Usage {#sec-deepspeed-usage}</span></span>
<span id="cb4-36"><a href="#cb4-36" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-37"><a href="#cb4-37" aria-hidden="true" tabindex="-1"></a><span class="in">```{.bash}</span></span>
<span id="cb4-38"><a href="#cb4-38" aria-hidden="true" tabindex="-1"></a><span class="in"># Passing arg via config</span></span>
<span id="cb4-39"><a href="#cb4-39" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl train config.yml</span></span>
<span id="cb4-40"><a href="#cb4-40" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-41"><a href="#cb4-41" aria-hidden="true" tabindex="-1"></a><span class="in"># Passing arg via cli</span></span>
<span id="cb4-42"><a href="#cb4-42" aria-hidden="true" tabindex="-1"></a><span class="in">axolotl train config.yml --deepspeed deepspeed_configs/zero1.json</span></span>
<span id="cb4-43"><a href="#cb4-43" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb4-44"><a href="#cb4-44" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-45"><a href="#cb4-45" aria-hidden="true" tabindex="-1"></a><span class="fu">### ZeRO Stages {#sec-zero-stages}</span></span>
<span id="cb4-46"><a href="#cb4-46" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-47"><a href="#cb4-47" aria-hidden="true" tabindex="-1"></a>We provide default configurations for:</span>
<span id="cb4-48"><a href="#cb4-48" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-49"><a href="#cb4-49" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 1 (<span class="in">`zero1.json`</span>)</span>
<span id="cb4-50"><a href="#cb4-50" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 2 (<span class="in">`zero2.json`</span>)</span>
<span id="cb4-51"><a href="#cb4-51" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>ZeRO Stage 3 (<span class="in">`zero3.json`</span>)</span>
<span id="cb4-52"><a href="#cb4-52" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-53"><a href="#cb4-53" aria-hidden="true" tabindex="-1"></a>Choose based on your memory requirements and performance needs.</span>
<span id="cb4-54"><a href="#cb4-54" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-55"><a href="#cb4-55" aria-hidden="true" tabindex="-1"></a><span class="fu">## FSDP {#sec-fsdp}</span></span>
<span id="cb4-56"><a href="#cb4-56" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-57"><a href="#cb4-57" aria-hidden="true" tabindex="-1"></a><span class="fu">### Basic FSDP Configuration {#sec-fsdp-config}</span></span>
<span id="cb4-58"><a href="#cb4-58" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-59"><a href="#cb4-59" aria-hidden="true" tabindex="-1"></a><span class="in">```{.yaml}</span></span>
<span id="cb4-60"><a href="#cb4-60" aria-hidden="true" tabindex="-1"></a><span class="in">fsdp:</span></span>
<span id="cb4-61"><a href="#cb4-61" aria-hidden="true" tabindex="-1"></a><span class="in"> - full_shard</span></span>
<span id="cb4-62"><a href="#cb4-62" aria-hidden="true" tabindex="-1"></a><span class="in"> - auto_wrap</span></span>
<span id="cb4-63"><a href="#cb4-63" aria-hidden="true" tabindex="-1"></a><span class="in">fsdp_config:</span></span>
<span id="cb4-64"><a href="#cb4-64" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_offload_params: true</span></span>
<span id="cb4-65"><a href="#cb4-65" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_state_dict_type: FULL_STATE_DICT</span></span>
<span id="cb4-66"><a href="#cb4-66" aria-hidden="true" tabindex="-1"></a><span class="in"> fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer</span></span>
<span id="cb4-67"><a href="#cb4-67" aria-hidden="true" tabindex="-1"></a><span class="in">```</span></span>
<span id="cb4-68"><a href="#cb4-68" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-69"><a href="#cb4-69" aria-hidden="true" tabindex="-1"></a><span class="fu">### FSDP + QLoRA {#sec-fsdp-qlora}</span></span>
<span id="cb4-70"><a href="#cb4-70" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-71"><a href="#cb4-71" aria-hidden="true" tabindex="-1"></a>For combining FSDP with QLoRA, see our <span class="co">[</span><span class="ot">dedicated guide</span><span class="co">](fsdp_qlora.qmd)</span>.</span>
<span id="cb4-72"><a href="#cb4-72" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-73"><a href="#cb4-73" aria-hidden="true" tabindex="-1"></a><span class="fu">## Performance Optimization {#sec-performance}</span></span>
<span id="cb4-74"><a href="#cb4-74" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-75"><a href="#cb4-75" aria-hidden="true" tabindex="-1"></a><span class="fu">### Liger Kernel Integration {#sec-liger}</span></span>
<span id="cb4-76"><a href="#cb4-76" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-77"><a href="#cb4-77" aria-hidden="true" tabindex="-1"></a>Please see <span class="co">[</span><span class="ot">docs</span><span class="co">](custom_integrations.qmd#liger)</span> for more info.</span>
<span id="cb4-78"><a href="#cb4-78" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-79"><a href="#cb4-79" aria-hidden="true" tabindex="-1"></a><span class="fu">## Troubleshooting {#sec-troubleshooting}</span></span>
<span id="cb4-80"><a href="#cb4-80" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-81"><a href="#cb4-81" aria-hidden="true" tabindex="-1"></a><span class="fu">### NCCL Issues {#sec-nccl}</span></span>
<span id="cb4-82"><a href="#cb4-82" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-83"><a href="#cb4-83" aria-hidden="true" tabindex="-1"></a>For NCCL-related problems, see our <span class="co">[</span><span class="ot">NCCL troubleshooting guide</span><span class="co">](nccl.qmd)</span>.</span>
<span id="cb4-84"><a href="#cb4-84" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-85"><a href="#cb4-85" aria-hidden="true" tabindex="-1"></a><span class="fu">### Common Problems {#sec-common-problems}</span></span>
<span id="cb4-86"><a href="#cb4-86" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-87"><a href="#cb4-87" aria-hidden="true" tabindex="-1"></a>::: {.panel-tabset}</span>
<span id="cb4-88"><a href="#cb4-88" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-89"><a href="#cb4-89" aria-hidden="true" tabindex="-1"></a><span class="fu">## Memory Issues</span></span>
<span id="cb4-90"><a href="#cb4-90" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-91"><a href="#cb4-91" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Reduce <span class="in">`micro_batch_size`</span></span>
<span id="cb4-92"><a href="#cb4-92" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Reduce <span class="in">`eval_batch_size`</span></span>
<span id="cb4-93"><a href="#cb4-93" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Adjust <span class="in">`gradient_accumulation_steps`</span></span>
<span id="cb4-94"><a href="#cb4-94" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Consider using a higher ZeRO stage</span>
<span id="cb4-95"><a href="#cb4-95" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-96"><a href="#cb4-96" aria-hidden="true" tabindex="-1"></a><span class="fu">## Training Instability</span></span>
<span id="cb4-97"><a href="#cb4-97" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-98"><a href="#cb4-98" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Start with DeepSpeed ZeRO-2</span>
<span id="cb4-99"><a href="#cb4-99" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Monitor loss values</span>
<span id="cb4-100"><a href="#cb4-100" aria-hidden="true" tabindex="-1"></a><span class="ss">- </span>Check learning rates</span>
<span id="cb4-101"><a href="#cb4-101" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-102"><a href="#cb4-102" aria-hidden="true" tabindex="-1"></a>:::</span>
<span id="cb4-103"><a href="#cb4-103" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb4-104"><a href="#cb4-104" aria-hidden="true" tabindex="-1"></a>For more detailed troubleshooting, see our <span class="co">[</span><span class="ot">debugging guide</span><span class="co">](debugging.qmd)</span>.</span></code><button title="Copy to Clipboard" class="code-copy-button" data-in-quarto-modal=""><i class="bi"></i></button></pre></div>
</div></div></div></div></div>
</div> <!-- /content -->

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multi-node.html">Multi Node</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/multi-node.html">Multi Node</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -338,7 +441,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multi-node.html">Multi Node</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/multi-node.html">Multi Node</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Multi Node</h1>
</div>
@@ -378,8 +481,8 @@ Important
<p>Make sure the main machine is reachable by other machines.</p>
</div>
</div>
<section id="accelerate" class="level1">
<h1>Accelerate</h1>
<section id="accelerate" class="level2">
<h2 class="anchored" data-anchor-id="accelerate">Accelerate</h2>
<p>You will need to create a configuration for accelerate, either by using <code>accelerate config</code> and follow the instructions or you can use one of the preset below:</p>
<p>~/.cache/huggingface/accelerate/default_config.yaml</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">compute_environment</span><span class="kw">:</span><span class="at"> LOCAL_MACHINE</span></span>
@@ -409,17 +512,17 @@ Important
<span id="cb2-7"><a href="#cb2-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">fsdp_transformer_layer_cls_to_wrap</span><span class="kw">:</span><span class="at"> LlamaDecoderLayer</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>All you have to do now is launch using accelerate as you would usually do on each machine and voila, the processes will start once you have launched accelerate on every machine.</p>
</section>
<section id="raytrain" class="level1">
<h1>Raytrain</h1>
<section id="raytrain" class="level2">
<h2 class="anchored" data-anchor-id="raytrain">Raytrain</h2>
<p>Please see ray train doc <a href="../docs/ray-integration.html">here</a>.</p>
</section>
<section id="torchrun" class="level1">
<h1>Torchrun</h1>
<section id="torchrun" class="level2">
<h2 class="anchored" data-anchor-id="torchrun">Torchrun</h2>
<p>If you are using Infiniband, we recommend torchrun to utilize the full bandwidth.</p>
<p>Set the following env (change buffersize/socketname depending on your system):</p>
<div class="sourceCode" id="cb3"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="at">export NCCL_IB_DISABLE=0</span></span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="at">export NCCL_SOCKET_IFNAME="eth0,en,eth,em,bond"</span></span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a><span class="at">export NCCL_BUFFSIZE=2097152</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_IB_DISABLE</span><span class="op">=</span>0</span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_SOCKET_IFNAME</span><span class="op">=</span><span class="st">"eth0,en,eth,em,bond"</span></span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_BUFFSIZE</span><span class="op">=</span>2097152</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Run the following on each node:</p>
<div class="sourceCode" id="cb4"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="ex">torchrun</span> <span class="at">--nnodes</span> <span class="va">$num_nodes</span> <span class="at">--nproc_per_node</span> <span class="va">$gpu_per_node</span> <span class="at">--rdzv_id</span> <span class="va">$rdzv_id</span> <span class="at">--rdzv_backend</span> c10d <span class="at">--rdzv_endpoint</span> <span class="st">"</span><span class="va">$head_node_ip</span><span class="st">:</span><span class="va">$head_node_port</span><span class="st">"</span> <span class="at">-m</span> axolotl.cli.train config.yaml</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Please make sure to substitute the placeholder variables.</p>

View File

@@ -69,10 +69,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -112,8 +112,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -129,7 +129,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multimodal.html">MultiModal / Vision Language Models (BETA)</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -150,7 +150,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -159,91 +159,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -298,7 +232,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -306,18 +240,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -337,7 +440,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
</nav>
</div>
<!-- main -->
<main class="content" id="quarto-document-content"><header id="title-block-header" class="quarto-title-block"></header>
<main class="content" id="quarto-document-content"><header id="title-block-header" class="quarto-title-block"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multimodal.html">MultiModal / Vision Language Models (BETA)</a></li></ol></nav></header>

View File

@@ -36,10 +36,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +79,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +96,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multipack.html">Multipack (Sample Packing)</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/multipack.html">Multipack (Sample Packing)</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +117,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +126,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +199,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +207,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -303,7 +406,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/multipack.html">Multipack (Sample Packing)</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/batch_vs_grad.html">Core Concepts</a></li><li class="breadcrumb-item"><a href="../docs/multipack.html">Multipack (Sample Packing)</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Multipack (Sample Packing)</h1>
</div>

View File

@@ -21,6 +21,40 @@ ul.task-list li input[type="checkbox"] {
margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */
vertical-align: middle;
}
/* CSS for syntax highlighting */
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
div.sourceCode { margin: 1em 0; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { display: inline-block; text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
{ counter-reset: source-line 0; }
pre.numberSource code > span
{ position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
{ content: counter(source-line);
position: relative; left: -1em; text-align: right; vertical-align: baseline;
border: none; display: inline-block;
-webkit-touch-callout: none; -webkit-user-select: none;
-khtml-user-select: none; -moz-user-select: none;
-ms-user-select: none; user-select: none;
padding: 0 4px; width: 4em;
}
pre.numberSource { margin-left: 3em; padding-left: 4px; }
div.sourceCode
{ }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
</style>
@@ -36,10 +70,10 @@ ul.task-list li input[type="checkbox"] {
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -79,8 +113,8 @@ ul.task-list li input[type="checkbox"] {
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -96,7 +130,7 @@ ul.task-list li input[type="checkbox"] {
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/nccl.html">NCCL</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/nccl.html">NCCL</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -117,7 +151,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -126,91 +160,25 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -265,7 +233,7 @@ ul.task-list li input[type="checkbox"] {
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -273,18 +241,187 @@ ul.task-list li input[type="checkbox"] {
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -296,7 +433,7 @@ ul.task-list li input[type="checkbox"] {
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/nccl.html">NCCL</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/faq.html">Troubleshooting</a></li><li class="breadcrumb-item"><a href="../docs/nccl.html">NCCL</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">NCCL</h1>
</div>
@@ -324,9 +461,9 @@ ul.task-list li input[type="checkbox"] {
<pre class="text"><code>Watchdog caught collective operation timeout: WorkNCCL(SeqNum=42, OpType=ALLGATHER, Timeout(ms)=1800000) ran for 1806948 milliseconds before timing out.</code></pre>
<p>Often, this timeout will happen after 30 minutes (the default setting) and is accompanied by below-average power consumption with near 100% GPU utilization before the error is raised. Nvidia recommends <a href="https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/troubleshooting.html#pci-access-control-services-acs">disabling PCI access control services (ACS)</a> as a possible solution if this is available to you.</p>
<p>Forcing cross-GPU communication via <a href="https://en.wikipedia.org/wiki/NVLink">NVLink</a> may help without increasing timeouts. To verify that your configuration is leveraging NVLink run the following command:</p>
<pre class="shell"><code>nvidia-smi nvlink --status</code></pre>
<div class="sourceCode" id="cb2"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="ex">nvidia-smi</span> nvlink <span class="at">--status</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>To force NCCL to use NVLink, simply set this in the environment:</p>
<pre class="shell"><code>export NCCL_P2P_LEVEL=NVL</code></pre>
<div class="sourceCode" id="cb3"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_P2P_LEVEL</span><span class="op">=</span>NVL</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>If NVLink is not available in your environment there are other options for <code>NCCL_P2P_LEVEL</code> in the table below:</p>
<table class="caption-top table">
<colgroup>
@@ -355,12 +492,12 @@ ul.task-list li input[type="checkbox"] {
</tbody>
</table>
<p>To validate that acceptable data transfer speeds exist for your training job, running <a href="https://github.com/NVIDIA/nccl-tests/blob/master/README.md">NCCL Tests</a> can help pinpoint bottlenecks, for example:</p>
<pre class="shell"><code>./build/all_reduce_perf -b 8 -e 128M -f 2 -g 3</code></pre>
<div class="sourceCode" id="cb4"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="ex">./build/all_reduce_perf</span> <span class="at">-b</span> 8 <span class="at">-e</span> 128M <span class="at">-f</span> 2 <span class="at">-g</span> 3</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>It can be useful when debugging NCCL communication timeouts to activate additional logging in both PyTorch and NCCL:</p>
<pre class="shell"><code>export NCCL_DEBUG=INFO
export NCCL_DEBUG_SUBSYS=ALL
export TORCH_DISTRIBUTED_DEBUG=INFO
export TORCHELASTIC_ERROR_FILE=/PATH/TO/torcherror.log</code></pre>
<div class="sourceCode" id="cb5"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_DEBUG</span><span class="op">=</span>INFO</span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">NCCL_DEBUG_SUBSYS</span><span class="op">=</span>ALL</span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">TORCH_DISTRIBUTED_DEBUG</span><span class="op">=</span>INFO</span>
<span id="cb5-4"><a href="#cb5-4" aria-hidden="true" tabindex="-1"></a><span class="bu">export</span> <span class="va">TORCHELASTIC_ERROR_FILE</span><span class="op">=</span>/PATH/TO/torcherror.log</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>Finally, if you believe your training job needs more time you can increase the timeout past 30 minutes by setting the <code>ddp_timeout</code> value in the Axolotl configuration. See <a href="https://pytorch.org/docs/stable/distributed.html#torch.distributed.init_process_group">PyTorch init_process_group</a> for documentation on this value.</p>

View File

@@ -8,7 +8,7 @@
<meta name="description" content="How to use Axolotl with Ray Train">
<title>Ray Train integration Axolotl</title>
<title>Ray Train Axolotl</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/ray-integration.html">Ray Train integration</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/ray-integration.html">Ray Train</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -339,9 +442,9 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/ray-integration.html">Ray Train integration</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multi-gpu.html">Deployments</a></li><li class="breadcrumb-item"><a href="../docs/ray-integration.html">Ray Train</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Ray Train integration</h1>
<h1 class="title">Ray Train</h1>
</div>
<div>
@@ -367,7 +470,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<p>With the <code>--use-ray</code> CLI flag, Axolotl will use Ray Trains <a href="https://docs.ray.io/en/latest/train/api/doc/ray.train.torch.TorchTrainer.html#ray.train.torch.TorchTrainer"><code>TorchTrainer</code></a> to run training.</p>
<section id="ray-cluster-setup" class="level2">
<h2 class="anchored" data-anchor-id="ray-cluster-setup">Ray cluster setup</h2>
<p>A prerequisite using the Ray Train integration is to setup a Ray cluster on your desired node(s). For a detailed guide on how you can get started with ray clusters, check the official Ray docs here: https://docs.ray.io/en/latest/cluster/getting-started.html</p>
<p>A prerequisite using the Ray Train integration is to setup a Ray cluster on your desired node(s). For a detailed guide on how you can get started with ray clusters, check the official Ray docs <a href="https://docs.ray.io/en/latest/cluster/getting-started.html">here</a>.</p>
<p>Every Ray cluster has one <em>head</em> node and a set of worker nodes. The head node is just like any other worker node, but it also runs certain special processes related to scheduling and orchestration. Ray-enabled scripts are run on the head node and depending on the resources (number of CPUs, GPUs, etc) they request, will be scheduled to run certain tasks on the worker nodes. For more on key concepts behind a Ray cluster, you can refer this <a href="https://docs.ray.io/en/latest/cluster/key-concepts.html#cluster-key-concepts">doc</a>.</p>
</section>
<section id="sanity-check" class="level2">
@@ -402,13 +505,11 @@ Demands:
<h2 class="anchored" data-anchor-id="configuring-training-with-ray-train">Configuring training with Ray Train</h2>
<p>You can find an example configuration at <code>configs/llama-3/lora-1b-ray.yaml</code>.</p>
<p>The key parameters to note here are:</p>
<div class="sourceCode" id="cb3"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="co">...</span></span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="co">use_ray: true</span></span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a><span class="co">ray_num_workers: 4</span></span>
<span id="cb3-4"><a href="#cb3-4" aria-hidden="true" tabindex="-1"></a><span class="co"># optional</span></span>
<span id="cb3-5"><a href="#cb3-5" aria-hidden="true" tabindex="-1"></a><span class="co">resources_per_worker:</span></span>
<span id="cb3-6"><a href="#cb3-6" aria-hidden="true" tabindex="-1"></a><span class="co"> GPU: 1</span></span>
<span id="cb3-7"><a href="#cb3-7" aria-hidden="true" tabindex="-1"></a><span class="co">...</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb3"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="fu">use_ray</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a><span class="fu">ray_num_workers</span><span class="kw">:</span><span class="at"> </span><span class="dv">4</span></span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a><span class="co"># optional</span></span>
<span id="cb3-4"><a href="#cb3-4" aria-hidden="true" tabindex="-1"></a><span class="fu">resources_per_worker</span><span class="kw">:</span></span>
<span id="cb3-5"><a href="#cb3-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">GPU</span><span class="kw">:</span><span class="at"> </span><span class="dv">1</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<ul>
<li><code>use_ray</code>: This is the flag that enables the Ray Train integration. You can either use the corresponding <code>--use-ray</code> flag in the CLI or set <code>use_ray</code> in the config file.</li>
<li><code>ray_num_workers</code>: This is the number of workers/GPUs to use for training.</li>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">Reward Modelling</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/reward_modelling.html">Reward Modelling</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -338,7 +441,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/reward_modelling.html">Reward Modelling</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Reward Modelling</h1>
</div>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/rlhf.html">RLHF (Beta)</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/rlhf.html">RLHF (Beta)</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -331,7 +434,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul>
<li><a href="#overview" id="toc-overview" class="nav-link active" data-scroll-target="#overview">Overview</a></li>
<li><a href="#rlhf-using-axolotl" id="toc-rlhf-using-axolotl" class="nav-link" data-scroll-target="#rlhf-using-axolotl">RLHF using Axolotl</a>
<ul class="collapse">
<ul>
<li><a href="#dpo" id="toc-dpo" class="nav-link" data-scroll-target="#dpo">DPO</a>
<ul class="collapse">
<li><a href="#chatml.argilla" id="toc-chatml.argilla" class="nav-link" data-scroll-target="#chatml.argilla">chatml.argilla</a></li>
@@ -369,6 +472,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li><a href="#llama3.ultra-1" id="toc-llama3.ultra-1" class="nav-link" data-scroll-target="#llama3.ultra-1">llama3.ultra</a></li>
<li><a href="#user_defined.default-1" id="toc-user_defined.default-1" class="nav-link" data-scroll-target="#user_defined.default-1">user_defined.default</a></li>
</ul></li>
<li><a href="#grpo" id="toc-grpo" class="nav-link" data-scroll-target="#grpo">GRPO</a></li>
<li><a href="#using-local-dataset-files" id="toc-using-local-dataset-files" class="nav-link" data-scroll-target="#using-local-dataset-files">Using local dataset files</a></li>
<li><a href="#trl-auto-unwrapping-for-peft" id="toc-trl-auto-unwrapping-for-peft" class="nav-link" data-scroll-target="#trl-auto-unwrapping-for-peft">TRL auto-unwrapping for PEFT</a></li>
</ul></li>
@@ -378,7 +482,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/rlhf.html">RLHF (Beta)</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/multimodal.html">How To Guides</a></li><li class="breadcrumb-item"><a href="../docs/rlhf.html">RLHF (Beta)</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">RLHF (Beta)</h1>
</div>
@@ -402,19 +506,19 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
</header>
<section id="overview" class="level1">
<h1>Overview</h1>
<section id="overview" class="level2">
<h2 class="anchored" data-anchor-id="overview">Overview</h2>
<p>Reinforcement Learning from Human Feedback is a method whereby a language model is optimized from data using human feedback. Various methods include, but not limited to:</p>
<ul>
<li>Proximal Policy Optimization (PPO) (not yet supported in axolotl)</li>
<li><a href="#dpo">Direct Preference Optimization (DPO)</a></li>
<li><a href="#ipo">Identity Preference Optimization (IPO)</a></li>
<li><a href="#kto">Kahneman-Tversky Optimization (KTO)</a></li>
<li><a href="#orpo">Odds Ratio Preference Optimization (ORPO)</a></li>
<li>Proximal Policy Optimization (PPO) (not yet supported in axolotl)</li>
</ul>
</section>
<section id="rlhf-using-axolotl" class="level1">
<h1>RLHF using Axolotl</h1>
<section id="rlhf-using-axolotl" class="level2">
<h2 class="anchored" data-anchor-id="rlhf-using-axolotl">RLHF using Axolotl</h2>
<div class="callout callout-style-default callout-important callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
@@ -442,8 +546,8 @@ Tip
<p>You can find what each method supports by going into <code>src/axolotl/prompt_strategies/{method}</code> where <code>{method}</code> is one of our supported methods. The <code>type:</code> can be retrieved from <code>{method}.{function_name}</code>.</p>
</div>
</div>
<section id="dpo" class="level2">
<h2 class="anchored" data-anchor-id="dpo">DPO</h2>
<section id="dpo" class="level3">
<h3 class="anchored" data-anchor-id="dpo">DPO</h3>
<p>Example config:</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> dpo</span></span>
<span id="cb1-2"><a href="#cb1-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
@@ -454,8 +558,8 @@ Tip
<span id="cb1-7"><a href="#cb1-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">split</span><span class="kw">:</span><span class="at"> train</span></span>
<span id="cb1-8"><a href="#cb1-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chatml</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>DPO supports the following types with the following dataset format:</p>
<section id="chatml.argilla" class="level3">
<h3 class="anchored" data-anchor-id="chatml.argilla">chatml.argilla</h3>
<section id="chatml.argilla" class="level4">
<h4 class="anchored" data-anchor-id="chatml.argilla">chatml.argilla</h4>
<div class="sourceCode" id="cb2"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb2-1"><a href="#cb2-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb2-2"><a href="#cb2-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb2-3"><a href="#cb2-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -463,8 +567,8 @@ Tip
<span id="cb2-5"><a href="#cb2-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected_response"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb2-6"><a href="#cb2-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.argilla_chat" class="level3">
<h3 class="anchored" data-anchor-id="chatml.argilla_chat">chatml.argilla_chat</h3>
<section id="chatml.argilla_chat" class="level4">
<h4 class="anchored" data-anchor-id="chatml.argilla_chat">chatml.argilla_chat</h4>
<div class="sourceCode" id="cb3"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"chosen"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span><span class="dt">"role"</span><span class="fu">:</span> <span class="st">"user"</span><span class="fu">,</span> <span class="dt">"content"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span><span class="ot">,</span></span>
@@ -476,8 +580,8 @@ Tip
<span id="cb3-9"><a href="#cb3-9" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb3-10"><a href="#cb3-10" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.icr" class="level3">
<h3 class="anchored" data-anchor-id="chatml.icr">chatml.icr</h3>
<section id="chatml.icr" class="level4">
<h4 class="anchored" data-anchor-id="chatml.icr">chatml.icr</h4>
<div class="sourceCode" id="cb4"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb4-2"><a href="#cb4-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb4-3"><a href="#cb4-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -485,8 +589,8 @@ Tip
<span id="cb4-5"><a href="#cb4-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb4-6"><a href="#cb4-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.intel" class="level3">
<h3 class="anchored" data-anchor-id="chatml.intel">chatml.intel</h3>
<section id="chatml.intel" class="level4">
<h4 class="anchored" data-anchor-id="chatml.intel">chatml.intel</h4>
<div class="sourceCode" id="cb5"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb5-2"><a href="#cb5-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb5-3"><a href="#cb5-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"question"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -494,8 +598,8 @@ Tip
<span id="cb5-5"><a href="#cb5-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb5-6"><a href="#cb5-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.prompt_pairs" class="level3">
<h3 class="anchored" data-anchor-id="chatml.prompt_pairs">chatml.prompt_pairs</h3>
<section id="chatml.prompt_pairs" class="level4">
<h4 class="anchored" data-anchor-id="chatml.prompt_pairs">chatml.prompt_pairs</h4>
<div class="sourceCode" id="cb6"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -503,8 +607,8 @@ Tip
<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb6-6"><a href="#cb6-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.ultra" class="level3">
<h3 class="anchored" data-anchor-id="chatml.ultra">chatml.ultra</h3>
<section id="chatml.ultra" class="level4">
<h4 class="anchored" data-anchor-id="chatml.ultra">chatml.ultra</h4>
<div class="sourceCode" id="cb7"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -518,8 +622,8 @@ Tip
<span id="cb7-11"><a href="#cb7-11" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb7-12"><a href="#cb7-12" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.argilla" class="level3">
<h3 class="anchored" data-anchor-id="llama3.argilla">llama3.argilla</h3>
<section id="llama3.argilla" class="level4">
<h4 class="anchored" data-anchor-id="llama3.argilla">llama3.argilla</h4>
<div class="sourceCode" id="cb8"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -527,8 +631,8 @@ Tip
<span id="cb8-5"><a href="#cb8-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected_response"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb8-6"><a href="#cb8-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.argilla_chat" class="level3">
<h3 class="anchored" data-anchor-id="llama3.argilla_chat">llama3.argilla_chat</h3>
<section id="llama3.argilla_chat" class="level4">
<h4 class="anchored" data-anchor-id="llama3.argilla_chat">llama3.argilla_chat</h4>
<div class="sourceCode" id="cb9"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb9-2"><a href="#cb9-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"chosen"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb9-3"><a href="#cb9-3" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span><span class="dt">"role"</span><span class="fu">:</span> <span class="st">"user"</span><span class="fu">,</span> <span class="dt">"content"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span><span class="ot">,</span></span>
@@ -540,8 +644,8 @@ Tip
<span id="cb9-9"><a href="#cb9-9" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb9-10"><a href="#cb9-10" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.icr" class="level3">
<h3 class="anchored" data-anchor-id="llama3.icr">llama3.icr</h3>
<section id="llama3.icr" class="level4">
<h4 class="anchored" data-anchor-id="llama3.icr">llama3.icr</h4>
<div class="sourceCode" id="cb10"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb10-2"><a href="#cb10-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb10-3"><a href="#cb10-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"input"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -549,8 +653,8 @@ Tip
<span id="cb10-5"><a href="#cb10-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb10-6"><a href="#cb10-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.intel" class="level3">
<h3 class="anchored" data-anchor-id="llama3.intel">llama3.intel</h3>
<section id="llama3.intel" class="level4">
<h4 class="anchored" data-anchor-id="llama3.intel">llama3.intel</h4>
<div class="sourceCode" id="cb11"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb11-1"><a href="#cb11-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb11-2"><a href="#cb11-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb11-3"><a href="#cb11-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"question"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -558,8 +662,8 @@ Tip
<span id="cb11-5"><a href="#cb11-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb11-6"><a href="#cb11-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.prompt_pairs" class="level3">
<h3 class="anchored" data-anchor-id="llama3.prompt_pairs">llama3.prompt_pairs</h3>
<section id="llama3.prompt_pairs" class="level4">
<h4 class="anchored" data-anchor-id="llama3.prompt_pairs">llama3.prompt_pairs</h4>
<div class="sourceCode" id="cb12"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb12-1"><a href="#cb12-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb12-2"><a href="#cb12-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb12-3"><a href="#cb12-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -567,8 +671,8 @@ Tip
<span id="cb12-5"><a href="#cb12-5" aria-hidden="true" tabindex="-1"></a> <span class="dt">"rejected"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb12-6"><a href="#cb12-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.ultra" class="level3">
<h3 class="anchored" data-anchor-id="llama3.ultra">llama3.ultra</h3>
<section id="llama3.ultra" class="level4">
<h4 class="anchored" data-anchor-id="llama3.ultra">llama3.ultra</h4>
<div class="sourceCode" id="cb13"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb13-1"><a href="#cb13-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb13-2"><a href="#cb13-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb13-3"><a href="#cb13-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
@@ -582,8 +686,8 @@ Tip
<span id="cb13-11"><a href="#cb13-11" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb13-12"><a href="#cb13-12" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="zephyr.nectar" class="level3">
<h3 class="anchored" data-anchor-id="zephyr.nectar">zephyr.nectar</h3>
<section id="zephyr.nectar" class="level4">
<h4 class="anchored" data-anchor-id="zephyr.nectar">zephyr.nectar</h4>
<div class="sourceCode" id="cb14"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb14-1"><a href="#cb14-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb14-2"><a href="#cb14-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb14-3"><a href="#cb14-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"answers"</span><span class="fu">:</span> <span class="ot">[</span></span>
@@ -599,8 +703,8 @@ Tip
<span id="cb14-13"><a href="#cb14-13" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb14-14"><a href="#cb14-14" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chat_template.default" class="level3">
<h3 class="anchored" data-anchor-id="chat_template.default">chat_template.default</h3>
<section id="chat_template.default" class="level4">
<h4 class="anchored" data-anchor-id="chat_template.default">chat_template.default</h4>
<div class="sourceCode" id="cb15"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb15-1"><a href="#cb15-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> dpo</span></span>
<span id="cb15-2"><a href="#cb15-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb15-3"><a href="#cb15-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> ...</span></span>
@@ -639,8 +743,8 @@ Tip
<span id="cb16-20"><a href="#cb16-20" aria-hidden="true" tabindex="-1"></a> <span class="fu">}</span></span>
<span id="cb16-21"><a href="#cb16-21" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="user_defined.default" class="level3">
<h3 class="anchored" data-anchor-id="user_defined.default">user_defined.default</h3>
<section id="user_defined.default" class="level4">
<h4 class="anchored" data-anchor-id="user_defined.default">user_defined.default</h4>
<p>For custom behaviors,</p>
<div class="sourceCode" id="cb17"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb17-1"><a href="#cb17-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> dpo</span></span>
<span id="cb17-2"><a href="#cb17-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
@@ -664,13 +768,13 @@ Tip
<span id="cb18-6"><a href="#cb18-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
<section id="ipo" class="level2">
<h2 class="anchored" data-anchor-id="ipo">IPO</h2>
<section id="ipo" class="level3">
<h3 class="anchored" data-anchor-id="ipo">IPO</h3>
<p>As IPO is just DPO with a different loss function, all supported options for DPO works here.</p>
<div class="sourceCode" id="cb19"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb19-1"><a href="#cb19-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> ipo</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="orpo" class="level2">
<h2 class="anchored" data-anchor-id="orpo">ORPO</h2>
<section id="orpo" class="level3">
<h3 class="anchored" data-anchor-id="orpo">ORPO</h3>
<p>Paper: https://arxiv.org/abs/2403.07691</p>
<div class="sourceCode" id="cb20"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb20-1"><a href="#cb20-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> orpo</span></span>
<span id="cb20-2"><a href="#cb20-2" aria-hidden="true" tabindex="-1"></a><span class="fu">orpo_alpha</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.1</span></span>
@@ -681,8 +785,8 @@ Tip
<span id="cb20-7"><a href="#cb20-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> argilla/ultrafeedback-binarized-preferences-cleaned</span></span>
<span id="cb20-8"><a href="#cb20-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chat_template.argilla</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>ORPO supports the following types with the following dataset format:</p>
<section id="chat_template.argilla" class="level3">
<h3 class="anchored" data-anchor-id="chat_template.argilla">chat_template.argilla</h3>
<section id="chat_template.argilla" class="level4">
<h4 class="anchored" data-anchor-id="chat_template.argilla">chat_template.argilla</h4>
<div class="sourceCode" id="cb21"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb21-1"><a href="#cb21-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb21-2"><a href="#cb21-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb21-3"><a href="#cb21-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">if</span> <span class="er">available</span><span class="fu">,</span> <span class="er">will</span> <span class="er">be</span> <span class="er">taken</span> <span class="er">as</span> <span class="er">user</span> <span class="er">message</span> <span class="er">for</span> <span class="er">single-turn</span> <span class="er">instead</span> <span class="er">of</span> <span class="er">from</span> <span class="er">list</span> <span class="er">below</span></span>
@@ -699,8 +803,8 @@ Tip
<span id="cb21-14"><a href="#cb21-14" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
<section id="kto" class="level2">
<h2 class="anchored" data-anchor-id="kto">KTO</h2>
<section id="kto" class="level3">
<h3 class="anchored" data-anchor-id="kto">KTO</h3>
<div class="sourceCode" id="cb22"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb22-1"><a href="#cb22-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> kto</span></span>
<span id="cb22-2"><a href="#cb22-2" aria-hidden="true" tabindex="-1"></a><span class="fu">rl_beta</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.5</span></span>
<span id="cb22-3"><a href="#cb22-3" aria-hidden="true" tabindex="-1"></a><span class="fu">kto_desirable_weight</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.2</span></span>
@@ -716,16 +820,16 @@ Tip
<span id="cb22-13"><a href="#cb22-13" aria-hidden="true" tabindex="-1"></a><span class="fu">gradient_checkpointing_kwargs</span><span class="kw">:</span></span>
<span id="cb22-14"><a href="#cb22-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">use_reentrant</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>KTO supports the following types with the following dataset format:</p>
<section id="chatml.argilla-1" class="level3">
<h3 class="anchored" data-anchor-id="chatml.argilla-1">chatml.argilla</h3>
<section id="chatml.argilla-1" class="level4">
<h4 class="anchored" data-anchor-id="chatml.argilla-1">chatml.argilla</h4>
<div class="sourceCode" id="cb23"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb23-1"><a href="#cb23-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb23-2"><a href="#cb23-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb23-3"><a href="#cb23-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb23-4"><a href="#cb23-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb23-5"><a href="#cb23-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.argilla_chat-1" class="level3">
<h3 class="anchored" data-anchor-id="chatml.argilla_chat-1">chatml.argilla_chat</h3>
<section id="chatml.argilla_chat-1" class="level4">
<h4 class="anchored" data-anchor-id="chatml.argilla_chat-1">chatml.argilla_chat</h4>
<div class="sourceCode" id="cb24"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb24-1"><a href="#cb24-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb24-2"><a href="#cb24-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"chosen"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb24-3"><a href="#cb24-3" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span><span class="dt">"role"</span><span class="fu">:</span> <span class="st">"user"</span><span class="fu">,</span> <span class="dt">"content"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span></span>
@@ -735,40 +839,40 @@ Tip
<span id="cb24-7"><a href="#cb24-7" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb24-8"><a href="#cb24-8" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.intel-1" class="level3">
<h3 class="anchored" data-anchor-id="chatml.intel-1">chatml.intel</h3>
<section id="chatml.intel-1" class="level4">
<h4 class="anchored" data-anchor-id="chatml.intel-1">chatml.intel</h4>
<div class="sourceCode" id="cb25"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb25-1"><a href="#cb25-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb25-2"><a href="#cb25-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb25-3"><a href="#cb25-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"question"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb25-4"><a href="#cb25-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb25-5"><a href="#cb25-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.prompt_pairs-1" class="level3">
<h3 class="anchored" data-anchor-id="chatml.prompt_pairs-1">chatml.prompt_pairs</h3>
<section id="chatml.prompt_pairs-1" class="level4">
<h4 class="anchored" data-anchor-id="chatml.prompt_pairs-1">chatml.prompt_pairs</h4>
<div class="sourceCode" id="cb26"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb26-1"><a href="#cb26-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb26-2"><a href="#cb26-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb26-3"><a href="#cb26-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb26-4"><a href="#cb26-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb26-5"><a href="#cb26-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="chatml.ultra-1" class="level3">
<h3 class="anchored" data-anchor-id="chatml.ultra-1">chatml.ultra</h3>
<section id="chatml.ultra-1" class="level4">
<h4 class="anchored" data-anchor-id="chatml.ultra-1">chatml.ultra</h4>
<div class="sourceCode" id="cb27"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb27-1"><a href="#cb27-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb27-2"><a href="#cb27-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb27-3"><a href="#cb27-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb27-4"><a href="#cb27-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb27-5"><a href="#cb27-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.argilla-1" class="level3">
<h3 class="anchored" data-anchor-id="llama3.argilla-1">llama3.argilla</h3>
<section id="llama3.argilla-1" class="level4">
<h4 class="anchored" data-anchor-id="llama3.argilla-1">llama3.argilla</h4>
<div class="sourceCode" id="cb28"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb28-1"><a href="#cb28-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb28-2"><a href="#cb28-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb28-3"><a href="#cb28-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"instruction"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb28-4"><a href="#cb28-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb28-5"><a href="#cb28-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.argilla_chat-1" class="level3">
<h3 class="anchored" data-anchor-id="llama3.argilla_chat-1">llama3.argilla_chat</h3>
<section id="llama3.argilla_chat-1" class="level4">
<h4 class="anchored" data-anchor-id="llama3.argilla_chat-1">llama3.argilla_chat</h4>
<div class="sourceCode" id="cb29"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb29-1"><a href="#cb29-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb29-2"><a href="#cb29-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="ot">[</span></span>
<span id="cb29-3"><a href="#cb29-3" aria-hidden="true" tabindex="-1"></a> <span class="fu">{</span><span class="dt">"role"</span><span class="fu">:</span> <span class="st">"user"</span><span class="fu">,</span> <span class="dt">"content"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">}</span><span class="ot">,</span></span>
@@ -776,32 +880,32 @@ Tip
<span id="cb29-5"><a href="#cb29-5" aria-hidden="true" tabindex="-1"></a> <span class="ot">]</span></span>
<span id="cb29-6"><a href="#cb29-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.intel-1" class="level3">
<h3 class="anchored" data-anchor-id="llama3.intel-1">llama3.intel</h3>
<section id="llama3.intel-1" class="level4">
<h4 class="anchored" data-anchor-id="llama3.intel-1">llama3.intel</h4>
<div class="sourceCode" id="cb30"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb30-1"><a href="#cb30-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb30-2"><a href="#cb30-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb30-3"><a href="#cb30-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"question"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb30-4"><a href="#cb30-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb30-5"><a href="#cb30-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.prompt_pairs-1" class="level3">
<h3 class="anchored" data-anchor-id="llama3.prompt_pairs-1">llama3.prompt_pairs</h3>
<section id="llama3.prompt_pairs-1" class="level4">
<h4 class="anchored" data-anchor-id="llama3.prompt_pairs-1">llama3.prompt_pairs</h4>
<div class="sourceCode" id="cb31"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb31-1"><a href="#cb31-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb31-2"><a href="#cb31-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb31-3"><a href="#cb31-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb31-4"><a href="#cb31-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb31-5"><a href="#cb31-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="llama3.ultra-1" class="level3">
<h3 class="anchored" data-anchor-id="llama3.ultra-1">llama3.ultra</h3>
<section id="llama3.ultra-1" class="level4">
<h4 class="anchored" data-anchor-id="llama3.ultra-1">llama3.ultra</h4>
<div class="sourceCode" id="cb32"><pre class="sourceCode json code-with-copy"><code class="sourceCode json"><span id="cb32-1"><a href="#cb32-1" aria-hidden="true" tabindex="-1"></a><span class="fu">{</span></span>
<span id="cb32-2"><a href="#cb32-2" aria-hidden="true" tabindex="-1"></a> <span class="dt">"system"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span> <span class="er">//</span> <span class="er">optional</span></span>
<span id="cb32-3"><a href="#cb32-3" aria-hidden="true" tabindex="-1"></a> <span class="dt">"prompt"</span><span class="fu">:</span> <span class="st">"..."</span><span class="fu">,</span></span>
<span id="cb32-4"><a href="#cb32-4" aria-hidden="true" tabindex="-1"></a> <span class="dt">"completion"</span><span class="fu">:</span> <span class="st">"..."</span></span>
<span id="cb32-5"><a href="#cb32-5" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="user_defined.default-1" class="level3">
<h3 class="anchored" data-anchor-id="user_defined.default-1">user_defined.default</h3>
<section id="user_defined.default-1" class="level4">
<h4 class="anchored" data-anchor-id="user_defined.default-1">user_defined.default</h4>
<p>For custom behaviors,</p>
<div class="sourceCode" id="cb33"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb33-1"><a href="#cb33-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> kto</span></span>
<span id="cb33-2"><a href="#cb33-2" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
@@ -824,20 +928,54 @@ Tip
<span id="cb34-6"><a href="#cb34-6" aria-hidden="true" tabindex="-1"></a><span class="fu">}</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
</section>
<section id="using-local-dataset-files" class="level2">
<h2 class="anchored" data-anchor-id="using-local-dataset-files">Using local dataset files</h2>
<div class="sourceCode" id="cb35"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb35-1"><a href="#cb35-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb35-2"><a href="#cb35-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">ds_type</span><span class="kw">:</span><span class="at"> json</span></span>
<span id="cb35-3"><a href="#cb35-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">data_files</span><span class="kw">:</span></span>
<span id="cb35-4"><a href="#cb35-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> orca_rlhf.jsonl</span></span>
<span id="cb35-5"><a href="#cb35-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">split</span><span class="kw">:</span><span class="at"> train</span></span>
<span id="cb35-6"><a href="#cb35-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chatml.intel</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<section id="grpo" class="level3">
<h3 class="anchored" data-anchor-id="grpo">GRPO</h3>
<p>GRPO uses custom reward functions and transformations. Please have them ready locally.</p>
<p>For ex, to load OpenAIs GSM8K and use a random reward for completions:</p>
<div class="sourceCode" id="cb35"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb35-1"><a href="#cb35-1" aria-hidden="true" tabindex="-1"></a><span class="co"># rewards.py</span></span>
<span id="cb35-2"><a href="#cb35-2" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> random</span>
<span id="cb35-3"><a href="#cb35-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb35-4"><a href="#cb35-4" aria-hidden="true" tabindex="-1"></a><span class="kw">def</span> rand_reward_func(completions, <span class="op">**</span>kwargs) <span class="op">-&gt;</span> <span class="bu">list</span>[<span class="bu">float</span>]:</span>
<span id="cb35-5"><a href="#cb35-5" aria-hidden="true" tabindex="-1"></a> <span class="cf">return</span> [random.uniform(<span class="dv">0</span>, <span class="dv">1</span>) <span class="cf">for</span> _ <span class="kw">in</span> completions]</span>
<span id="cb35-6"><a href="#cb35-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb35-7"><a href="#cb35-7" aria-hidden="true" tabindex="-1"></a><span class="kw">def</span> oai_gsm8k_transform(cfg, <span class="op">*</span>args, <span class="op">**</span>kwargs):</span>
<span id="cb35-8"><a href="#cb35-8" aria-hidden="true" tabindex="-1"></a> <span class="kw">def</span> transform_fn(example, tokenizer<span class="op">=</span><span class="va">None</span>):</span>
<span id="cb35-9"><a href="#cb35-9" aria-hidden="true" tabindex="-1"></a> label <span class="op">=</span> example[<span class="st">"answer"</span>].split(<span class="st">"####"</span>)[<span class="op">-</span><span class="dv">1</span>].strip().replace(<span class="st">","</span>, <span class="st">""</span>)</span>
<span id="cb35-10"><a href="#cb35-10" aria-hidden="true" tabindex="-1"></a> <span class="cf">return</span> {</span>
<span id="cb35-11"><a href="#cb35-11" aria-hidden="true" tabindex="-1"></a> <span class="st">"prompt"</span>: [{<span class="st">"role"</span>: <span class="st">"user"</span>, <span class="st">"content"</span>: example[<span class="st">"question"</span>]},],</span>
<span id="cb35-12"><a href="#cb35-12" aria-hidden="true" tabindex="-1"></a> <span class="st">"answer"</span>: label,</span>
<span id="cb35-13"><a href="#cb35-13" aria-hidden="true" tabindex="-1"></a> }</span>
<span id="cb35-14"><a href="#cb35-14" aria-hidden="true" tabindex="-1"></a> <span class="cf">return</span> transform_fn, {<span class="st">"remove_columns"</span>: [<span class="st">"question"</span>]}</span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb36"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb36-1"><a href="#cb36-1" aria-hidden="true" tabindex="-1"></a><span class="fu">rl</span><span class="kw">:</span><span class="at"> grpo</span></span>
<span id="cb36-2"><a href="#cb36-2" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb36-3"><a href="#cb36-3" aria-hidden="true" tabindex="-1"></a><span class="fu">trl</span><span class="kw">:</span></span>
<span id="cb36-4"><a href="#cb36-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">beta</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.001</span></span>
<span id="cb36-5"><a href="#cb36-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">max_completion_length</span><span class="kw">:</span><span class="at"> </span><span class="dv">256</span></span>
<span id="cb36-6"><a href="#cb36-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">use_vllm</span><span class="kw">:</span><span class="at"> </span><span class="ch">True</span></span>
<span id="cb36-7"><a href="#cb36-7" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">vllm_device</span><span class="kw">:</span><span class="at"> auto</span></span>
<span id="cb36-8"><a href="#cb36-8" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">vllm_gpu_memory_utilization</span><span class="kw">:</span><span class="at"> </span><span class="fl">0.15</span></span>
<span id="cb36-9"><a href="#cb36-9" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">num_generations</span><span class="kw">:</span><span class="at"> </span><span class="dv">4</span></span>
<span id="cb36-10"><a href="#cb36-10" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">reward_funcs</span><span class="kw">:</span><span class="at"> </span><span class="kw">[</span><span class="st">"rewards.rand_reward_func"</span><span class="kw">]</span><span class="co"> # format: '{file_name}.{fn_name}'</span></span>
<span id="cb36-11"><a href="#cb36-11" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb36-12"><a href="#cb36-12" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">path</span><span class="kw">:</span><span class="at"> openai/gsm8k</span></span>
<span id="cb36-13"><a href="#cb36-13" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">name</span><span class="kw">:</span><span class="at"> main</span></span>
<span id="cb36-14"><a href="#cb36-14" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> rewards.oai_gsm8k_transform</span><span class="co"> # format: '{file_name}.{fn_name}'</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<p>To see other examples of custom reward functions, please see <a href="https://github.com/huggingface/trl/blob/main/docs/source/grpo_trainer.md#using-a-custom-reward-function">TRL GRPO Docs</a>.</p>
</section>
<section id="trl-auto-unwrapping-for-peft" class="level2">
<h2 class="anchored" data-anchor-id="trl-auto-unwrapping-for-peft">TRL auto-unwrapping for PEFT</h2>
<section id="using-local-dataset-files" class="level3">
<h3 class="anchored" data-anchor-id="using-local-dataset-files">Using local dataset files</h3>
<div class="sourceCode" id="cb37"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb37-1"><a href="#cb37-1" aria-hidden="true" tabindex="-1"></a><span class="fu">datasets</span><span class="kw">:</span></span>
<span id="cb37-2"><a href="#cb37-2" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> </span><span class="fu">ds_type</span><span class="kw">:</span><span class="at"> json</span></span>
<span id="cb37-3"><a href="#cb37-3" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">data_files</span><span class="kw">:</span></span>
<span id="cb37-4"><a href="#cb37-4" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="kw">-</span><span class="at"> orca_rlhf.jsonl</span></span>
<span id="cb37-5"><a href="#cb37-5" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">split</span><span class="kw">:</span><span class="at"> train</span></span>
<span id="cb37-6"><a href="#cb37-6" aria-hidden="true" tabindex="-1"></a><span class="at"> </span><span class="fu">type</span><span class="kw">:</span><span class="at"> chatml.intel</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="trl-auto-unwrapping-for-peft" class="level3">
<h3 class="anchored" data-anchor-id="trl-auto-unwrapping-for-peft">TRL auto-unwrapping for PEFT</h3>
<p>TRL supports auto-unwrapping PEFT models for RL training paradigms which rely on a reference model. This significantly reduces memory pressure as an additional refreference model does not need to be loaded, and reference model log-probabilities can be obtained by disabling PEFT adapters. This is enabled by default. To turn it off, pass the following config:</p>
<div class="sourceCode" id="cb36"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb36-1"><a href="#cb36-1" aria-hidden="true" tabindex="-1"></a><span class="co"># load ref model when adapter training.</span></span>
<span id="cb36-2"><a href="#cb36-2" aria-hidden="true" tabindex="-1"></a><span class="fu">rl_adapter_ref_model</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
<div class="sourceCode" id="cb38"><pre class="sourceCode yaml code-with-copy"><code class="sourceCode yaml"><span id="cb38-1"><a href="#cb38-1" aria-hidden="true" tabindex="-1"></a><span class="co"># load ref model when adapter training.</span></span>
<span id="cb38-2"><a href="#cb38-2" aria-hidden="true" tabindex="-1"></a><span class="fu">rl_adapter_ref_model</span><span class="kw">:</span><span class="at"> </span><span class="ch">true</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item">PyTorch ao</li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/torchao.html">PyTorch ao</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -336,7 +439,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/torchao.html">PyTorch ao</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">PyTorch ao</h1>
</div>
@@ -360,6 +463,20 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
</header>
<p>To use experimental optimizers (<code>AdamWFp8</code>, <code>AdamW4bit</code>, <code>AdamW8bit</code>) from Pytorch Ao, please install the package as shown below.</p>
<div class="callout callout-style-default callout-tip callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Tip
</div>
</div>
<div class="callout-body-container callout-body">
<p>Some experimental optimizers are already present in regular Pytorch, so please re-check if you actually need this package!</p>
</div>
</div>
<section id="installation" class="level3">
<h3 class="anchored" data-anchor-id="installation">Installation</h3>
<p>Stable Release from the PyTorch index</p>

View File

@@ -70,10 +70,10 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<script src="../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../site_libs/quarto-html/anchor.min.js"></script>
<link href="../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-2f5df379a58b258e96c21c0638c20c03.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<link href="../site_libs/quarto-html/quarto-syntax-highlighting-dark-b53751a350365c71b6c909e95f209ed1.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../site_libs/bootstrap/bootstrap-141b2cdb37a94fcfd6825c1581ff795f.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<link href="../site_libs/bootstrap/bootstrap-0cda210ced8960466d2ee7bf22d15016.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="dark">
<script id="quarto-search-options" type="application/json">{
"location": "navbar",
"copy-button": false,
@@ -113,8 +113,8 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<nav class="navbar navbar-expand " data-bs-theme="dark">
<div class="navbar-container container-fluid">
<div class="navbar-brand-container mx-auto">
<a class="navbar-brand" href="../index.html">
<span class="navbar-title">Axolotl</span>
<a href="../index.html" class="navbar-brand navbar-brand-logo">
<img src="../image/axolotl_logo_digital_white.svg" alt="" class="navbar-logo">
</a>
</div>
<div class="quarto-navbar-tools tools-wide tools-end">
@@ -130,7 +130,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
<i class="bi bi-layout-text-sidebar-reverse"></i>
</button>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/unsloth.html">Unsloth</a></li></ol></nav>
<nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/unsloth.html">Unsloth</a></li></ol></nav>
<a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
</a>
</div>
@@ -151,7 +151,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true">
<span class="menu-text">How-To Guides</span></a>
<span class="menu-text">Getting Started</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-1" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -160,91 +160,25 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/getting-started.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Getting Started with Axolotl</span></a>
<span class="menu-text">Quickstart</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/installation.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Installation Guide</span></a>
<span class="menu-text">Installation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
<a href="../docs/cli.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">CLI Reference</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/inference.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Inference Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/input_output.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Template-free prompt construction</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU Training Guide</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Training with AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train integration</span></a>
<span class="menu-text">Inference</span></a>
</div>
</li>
</ul>
@@ -299,7 +233,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<span class="menu-text">Deployments</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
@@ -307,18 +241,187 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-gpu.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi-GPU</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multi-node.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multi Node</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/ray-integration.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Ray Train</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/amd_hpc.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">AMD GPUs on HPC Systems</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/mac.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Mac M-series</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true">
<span class="menu-text">How To Guides</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-4" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-4" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multimodal.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">MultiModal / Vision Language Models (BETA)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/rlhf.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">RLHF (Beta)</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/reward_modelling.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Reward Modelling</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lr_groups.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Learning Rate Groups</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/lora_optims.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">LoRA Optimizations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true">
<span class="menu-text">Core Concepts</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-5" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-5" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/batch_vs_grad.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Batch size vs Gradient accumulation</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/dataset_preprocessing.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Dataset Preprocessing</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/multipack.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Multipack (Sample Packing)</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true">
<span class="menu-text">Advanced Features</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-6" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-6" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/fsdp_qlora.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FDSP + QLoRA</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/unsloth.html" class="sidebar-item-text sidebar-link active">
<span class="menu-text">Unsloth</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/torchao.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">PyTorch ao</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/custom_integrations.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Custom Integrations</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true">
<span class="menu-text">Troubleshooting</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-7" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-7" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/debugging.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Debugging</span></a>
</div>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/nccl.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">NCCL</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item sidebar-item-section">
<div class="sidebar-item-container">
<a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true">
<span class="menu-text">Reference</span></a>
<a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-8" role="navigation" aria-expanded="true" aria-label="Toggle section">
<i class="bi bi-chevron-right ms-2"></i>
</a>
</div>
<ul id="quarto-sidebar-section-8" class="collapse list-unstyled sidebar-section depth1 show">
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/config.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">Config options</span></a>
</div>
</li>
</ul>
</li>
<li class="sidebar-item">
<div class="sidebar-item-container">
<a href="../docs/faq.html" class="sidebar-item-text sidebar-link">
<span class="menu-text">FAQ</span></a>
</div>
</li>
</ul>
</div>
</nav>
@@ -331,7 +434,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<ul>
<li><a href="#overview" id="toc-overview" class="nav-link active" data-scroll-target="#overview">Overview</a></li>
<li><a href="#installation" id="toc-installation" class="nav-link" data-scroll-target="#installation">Installation</a></li>
<li><a href="#using-unsloth-w-axolotl" id="toc-using-unsloth-w-axolotl" class="nav-link" data-scroll-target="#using-unsloth-w-axolotl">Using unsloth w Axolotl</a></li>
<li><a href="#usage" id="toc-usage" class="nav-link" data-scroll-target="#usage">Usage</a></li>
<li><a href="#limitations" id="toc-limitations" class="nav-link" data-scroll-target="#limitations">Limitations</a></li>
</ul>
</nav>
@@ -339,7 +442,7 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<!-- main -->
<main class="content" id="quarto-document-content">
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/getting-started.html">How-To Guides</a></li><li class="breadcrumb-item"><a href="../docs/unsloth.html">Unsloth</a></li></ol></nav>
<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../docs/fsdp_qlora.html">Advanced Features</a></li><li class="breadcrumb-item"><a href="../docs/unsloth.html">Unsloth</a></li></ol></nav>
<div class="quarto-title">
<h1 class="title">Unsloth</h1>
</div>
@@ -366,14 +469,28 @@ pre > code.sourceCode > span > a:first-child::before { text-decoration: underlin
<section id="overview" class="level3">
<h3 class="anchored" data-anchor-id="overview">Overview</h3>
<p>Unsloth provides hand-written optimized kernels for LLM finetuning that slightly improve speed and VRAM over standard industry baselines.</p>
<div class="callout callout-style-default callout-important callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Important
</div>
</div>
<div class="callout-body-container callout-body">
<p>Due to breaking changes in transformers <code>v4.48.0</code>, users will need to downgrade to <code>&lt;=v4.47.1</code> to use this patch.</p>
<p>This will later be deprecated in favor of <a href="../docs/lora_optims.html">LoRA Optimizations</a>.</p>
</div>
</div>
</section>
<section id="installation" class="level3">
<h3 class="anchored" data-anchor-id="installation">Installation</h3>
<p>The following will install the correct unsloth and extras from source.</p>
<div class="sourceCode" id="cb1"><pre class="sourceCode bash code-with-copy"><code class="sourceCode bash"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="ex">python</span> scripts/unsloth_install.py <span class="kw">|</span> <span class="fu">sh</span></span></code><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></pre></div>
</section>
<section id="using-unsloth-w-axolotl" class="level3">
<h3 class="anchored" data-anchor-id="using-unsloth-w-axolotl">Using unsloth w Axolotl</h3>
<section id="usage" class="level3">
<h3 class="anchored" data-anchor-id="usage">Usage</h3>
<p>Axolotl exposes a few configuration options to try out unsloth and get most of the performance gains.</p>
<p>Our unsloth integration is currently limited to the following model architectures: - llama</p>
<p>These options are specific to LoRA finetuning and cannot be used for multi-GPU finetuning</p>