fix softmax class check

register rala
use autoconfig w rala
2025-01-15 23:23:13 -05:00 · 2025-01-15 23:21:22 -05:00 · 2025-01-15 23:14:47 -05:00 · 2025-01-15 22:45:02 -05:00 · 2025-01-15 21:36:14 -05:00 · 2025-01-15 21:27:12 -05:00
778 changed files with 20072 additions and 79779 deletions
--- a/.axolotl-complete.bash
+++ b/.axolotl-complete.bash
@@ -1,41 +0,0 @@
-#!/bin/bash
-
-_axolotl_completions() {
-    local cur prev
-    COMPREPLY=()
-    cur="${COMP_WORDS[COMP_CWORD]}"
-    prev="${COMP_WORDS[COMP_CWORD-1]}"
-
-    # If we're completing the first argument (the command)
-    if [[ $COMP_CWORD -eq 1 ]]; then
-        mapfile -t COMPREPLY < <(compgen -W "delinearize-llama4 fetch lm-eval merge-sharded-fsdp-weights quantize vllm-serve evaluate inference merge-lora preprocess train" -- "$cur")
-        return 0
-    fi
-
-    # Commands that should complete with directories and YAML files
-    local -a yaml_commands=("merge-sharded-fsdp-weights" "quantize" "vllm-serve" "evaluate" "inference" "merge-lora" "preprocess" "train")
-
-    # Check if previous word is in our list
-    if [[ " ${yaml_commands[*]} " =~ (^|[[:space:]])$prev($|[[:space:]]) ]]; then
-        # Use filename completion which handles directories properly
-        compopt -o filenames
-        mapfile -t COMPREPLY < <(compgen -f -- "$cur")
-
-        # Filter to only include directories and YAML files
-        local -a filtered=()
-        for item in "${COMPREPLY[@]}"; do
-            if [[ -d "$item" ]] || [[ "$item" == *.yaml ]] || [[ "$item" == *.yml ]]; then
-                filtered+=("$item")
-            fi
-        done
-        COMPREPLY=("${filtered[@]}")
-
-        return 0
-    fi
-
-    # Default: no completion
-    return 0
-}
-
-# Remove the -o nospace option - let filenames handle it
-complete -F _axolotl_completions axolotl
--- a/.bandit
+++ b/.bandit
@@ -1,3 +1,3 @@
 [bandit]
 exclude = tests
-skips = B101,B615
+skips = B101
--- a/.coderabbit.yaml
+++ b/.coderabbit.yaml
@@ -1,16 +0,0 @@
-# yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json
-language: "en-US"
-early_access: false
-reviews:
-  profile: "chill"
-  request_changes_workflow: false
-  high_level_summary: true
-  review_status: true
-  collapse_walkthrough: true
-  poem: false
-  sequence_diagrams: false
-  auto_review:
-    enabled: true
-    drafts: false
-chat:
-  auto_reply: true
--- a/.coveragerc
+++ b/.coveragerc
@@ -1,14 +0,0 @@
-[run]
-source = axolotl
-omit =
-    */tests/*
-    setup.py
-
-[report]
-exclude_lines =
-    pragma: no cover
-    def __repr__
-    raise NotImplementedError
-    if __name__ == .__main__.:
-    pass
-    raise ImportError
--- a/.github/CONTRIBUTING.md
+++ b/.github/CONTRIBUTING.md
@@ -15,7 +15,7 @@ First of all, thank you for your interest in contributing to axolotl! We appreci
  - [Commit Messages](#commit-messages)
 - [Additional Resources](#additional-resources)

-## Code of Conduct
+## Code of Conductcode

 All contributors are expected to adhere to our [Code of Conduct](CODE_OF_CONDUCT.md). Please read it before participating in the axolotl community.

@@ -57,13 +57,6 @@ We welcome ideas for improvements and new features. To suggest an enhancement, o
 5. Push your branch to your fork on GitHub.
 6. Open a new pull request against the `main` branch of the axolotl repository. Include a clear and concise description of your changes, referencing any related issues.

-#### Skipping CI Checks
-
-You can skip certain CI checks by including specific keywords in your commit messages:
-
- `[skip ci]` or `skip ci` - Skips all CI checks for that commit
- `[skip-e2e]` or `skip-e2e` - Skips only end-to-end tests while running other CI checks. You may also include this in the title of your PR to disable end-to-end tests for the entire PR.
-
 ## Style Guidelines

 ### Code Style
--- a/.github/workflows/base.yml
+++ b/.github/workflows/base.yml
@@ -5,83 +5,53 @@ on:
    branches:
      - "main"
    paths:
-      - 'docker/Dockerfile-base'
-      - 'docker/Dockerfile-uv-base'
+      - 'Dockerfile-base'
      - '.github/workflows/base.yml'
  pull_request:
    paths:
-      - 'docker/Dockerfile-base'
-      - 'docker/Dockerfile-uv-base'
+      - 'Dockerfile-base'
      - '.github/workflows/base.yml'
  workflow_dispatch:

 jobs:
  build-base:
-    if: ${{ github.repository_owner == 'axolotl-ai-cloud' && (github.event_name != 'pull_request' || !github.event.pull_request.draft) }}
-    timeout-minutes: 480
+    if: github.repository_owner == 'axolotl-ai-cloud'
    # this job needs to be run on self-hosted GPU runners...
-    runs-on: ubuntu-latest-m
+    runs-on: axolotl-gpu-runner
    strategy:
      fail-fast: false
      matrix:
        include:
+          - cuda: "121"
+            cuda_version: 12.1.1
+            cudnn_version: 8
+            python_version: "3.10"
+            pytorch: 2.3.1
+            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
+          - cuda: "121"
+            cuda_version: 12.1.1
+            cudnn_version: 8
+            python_version: "3.11"
+            pytorch: 2.3.1
+            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
+          - cuda: "124"
+            cuda_version: 12.4.1
+            cudnn_version: ""
+            python_version: "3.10"
+            pytorch: 2.4.1
+            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
          - cuda: "124"
            cuda_version: 12.4.1
            cudnn_version: ""
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.4.1
            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-          - cuda: "126"
-            cuda_version: 12.6.3
+          - cuda: "124"
+            cuda_version: 12.4.1
            cudnn_version: ""
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.5.1
            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-          - cuda: "126"
-            cuda_version: 12.6.3
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.7.0
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-          - cuda: "126"
-            cuda_version: 12.6.3
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.7.1
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-          - cuda: "128"
-            cuda_version: 12.8.1
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.7.1
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-          - cuda: "128"
-            cuda_version: 12.8.1
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.8.0
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-base"
-#          - cuda: "128"
-#            cuda_version: 12.8.1
-#            cudnn_version: ""
-#            python_version: "3.11"
-#            pytorch: nightly
-#            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-#            dockerfile: "Dockerfile-base-nightly"
-#          # "next" is for release candidates of pytorch
-#          - cuda: "128"
-#            cuda_version: 12.8.1
-#            cudnn_version: ""
-#            python_version: "3.11"
-#            pytorch: next
-#            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-#            dockerfile: "Dockerfile-base-next"
    steps:
      - name: Checkout
        uses: actions/checkout@v4
@@ -103,74 +73,7 @@ jobs:
        uses: docker/build-push-action@v4
        with:
          context: .
-          file: ./docker/${{ matrix.dockerfile }}
-          push: ${{ github.event_name != 'pull_request' }}
-          tags: ${{ steps.metadata.outputs.tags }}-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}${{ matrix.axolotl_extras != '' && '-' || '' }}${{ matrix.axolotl_extras }}
-          labels: ${{ steps.metadata.outputs.labels }}
-          build-args: |
-            CUDA_VERSION=${{ matrix.cuda_version }}
-            CUDNN_VERSION=${{ matrix.cudnn_version }}
-            CUDA=${{ matrix.cuda }}
-            PYTHON_VERSION=${{ matrix.python_version }}
-            PYTORCH_VERSION=${{ matrix.pytorch }}
-            TORCH_CUDA_ARCH_LIST=${{ matrix.torch_cuda_arch_list }}
-  build-base-uv:
-    if: ${{ github.repository_owner == 'axolotl-ai-cloud' && (github.event_name != 'pull_request' || !github.event.pull_request.draft) }}
-    timeout-minutes: 480
-    runs-on: ubuntu-latest-m
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - cuda: "126"
-            cuda_version: 12.6.3
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.6.0
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-uv-base"
-          - cuda: "126"
-            cuda_version: 12.6.3
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.7.1
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-uv-base"
-          - cuda: "128"
-            cuda_version: 12.8.1
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.7.1
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-uv-base"
-          - cuda: "128"
-            cuda_version: 12.8.1
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.8.0
-            torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
-            dockerfile: "Dockerfile-uv-base"
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Docker metadata
-        id: metadata
-        uses: docker/metadata-action@v5
-        with:
-          images: |
-            axolotlai/axolotl-base-uv
-      - name: Login to Docker Hub
-        uses: docker/login-action@v2
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
-      - name: Build
-        uses: docker/build-push-action@v4
-        with:
-          context: .
-          file: ./docker/${{ matrix.dockerfile }}
+          file: ./docker/Dockerfile-base
          push: ${{ github.event_name != 'pull_request' }}
          tags: ${{ steps.metadata.outputs.tags }}-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}${{ matrix.axolotl_extras != '' && '-' || '' }}${{ matrix.axolotl_extras }}
          labels: ${{ steps.metadata.outputs.labels }}
--- a/.github/workflows/docs.yml
+++ b/.github/workflows/docs.yml
@@ -19,13 +19,10 @@ jobs:
        - name: Setup Python
          uses: actions/setup-python@v5
          with:
-            python-version: '3.11'
-        - name: Install dependencies
+            python-version: '3.10'
+        - name: install dependencies
          run: |
-            python3 -m pip install jupyter quartodoc
-            python3 -m pip install -e .
-        - name: Build autodoc
-          run: quartodoc build
+            python3 -m pip install jupyter
        - name: Publish to GitHub Pages (and render)
          uses: quarto-dev/quarto-actions/publish@v2
          with:
--- a/.github/workflows/lint.yml
+++ b/.github/workflows/lint.yml
@@ -3,25 +3,22 @@ on:
  # check on PRs, and manual triggers
  merge_group:
  pull_request:
-      types: [opened, synchronize, reopened, ready_for_review]
      paths:
       - '**.py'
       - 'requirements.txt'
       - '.github/workflows/*.yml'
       - "*.[q]md"
       - "examples/**/*.y[a]?ml"
-       - ".pre-commit-config.yaml"
  workflow_dispatch:

 jobs:
  pre-commit:
    name: pre-commit
    runs-on: ubuntu-latest
-    if: ${{ !github.event.pull_request.draft }}
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
          cache: 'pip' # caching pip dependencies
      - uses: pre-commit/action@v3.0.1
--- a/.github/workflows/main.yml
+++ b/.github/workflows/main.yml
@@ -15,27 +15,27 @@ jobs:
      fail-fast: false
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
+            axolotl_extras: mamba-ssm
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.3.1
+            axolotl_extras: mamba-ssm
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.4.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.7.0
+            pytorch: 2.5.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras: vllm
            is_latest: true
-          - cuda: 128
-            cuda_version: 12.8.1
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras:
    runs-on: axolotl-gpu-runner
    steps:
      - name: Checkout
@@ -67,7 +67,6 @@ jobs:
            CUDA=${{ matrix.cuda }}
            PYTORCH_VERSION=${{ matrix.pytorch }}
            AXOLOTL_ARGS=${{ matrix.axolotl_args }}
-            AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}
          file: ./docker/Dockerfile
          push: ${{ github.event_name != 'pull_request' }}
          tags: |
@@ -83,33 +82,27 @@ jobs:
    strategy:
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.6.0
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.7.0
+            pytorch: 2.3.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.7.1
+            pytorch: 2.4.1
            axolotl_extras:
-            is_latest:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras: vllm
+            pytorch: 2.5.1
+            axolotl_extras:
            is_latest: true
-          - cuda: 128
-            cuda_version: 12.8.1
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras:
    runs-on: axolotl-gpu-runner
    steps:
      - name: Checkout
@@ -152,23 +145,11 @@ jobs:
    strategy:
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.3.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras:
-            is_latest:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras: vllm
-            is_latest: true
    runs-on: axolotl-gpu-runner
    steps:
      - name: Checkout
--- a/.github/workflows/multi-gpu-e2e.yml
+++ b/.github/workflows/multi-gpu-e2e.yml
@@ -3,13 +3,7 @@ name: docker-multigpu-tests-biweekly
 on:
  pull_request:
    paths:
-      - 'tests/e2e/multigpu/**.py'
-      - 'requirements.txt'
-      - 'setup.py'
-      - 'pyproject.toml'
-      - '.github/workflows/multi-gpu-e2e.yml'
-      - 'src/axolotl/core/trainers/mixins/sequence_parallel.py'
-      - 'src/axolotl/utils/distributed.py'
+      - 'tests/e2e/multigpu/*.py'
  workflow_dispatch:
  schedule:
    - cron: '0 0 * * 1,4'  # Runs at 00:00 UTC every monday & thursday
@@ -21,32 +15,31 @@ concurrency:

 jobs:
  test-axolotl-multigpu:
-    if: ${{ ! contains(github.event.commits[0].message, '[skip e2e]') && github.repository_owner == 'axolotl-ai-cloud' && (github.event_name != 'pull_request' || !github.event.pull_request.draft) }}
+    if: ${{ ! contains(github.event.commits[0].message, '[skip e2e]') && github.repository_owner == 'axolotl-ai-cloud' }}
    strategy:
      fail-fast: false
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.3.1
+            axolotl_extras:
+            num_gpus: 2
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.4.1
            axolotl_extras:
            num_gpus: 2
            nightly_build: "true"
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.7.0
+            pytorch: 2.5.1
            axolotl_extras:
            num_gpus: 2
            nightly_build: "true"
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            axolotl_extras: vllm
-            num_gpus: 2
-            nightly_build: "true"
    runs-on: [self-hosted, modal]
    timeout-minutes: 120
    steps:
@@ -55,11 +48,11 @@ jobs:
      - name: Install Python
        uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
      - name: Install Modal
        run: |
          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
+          pip install modal==0.71.8 jinja2
      - name: Update env vars
        run: |
          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
@@ -69,7 +62,6 @@ jobs:
          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
          echo "NIGHTLY_BUILD=${{ matrix.nightly_build }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
      - name: Run tests job on Modal
        run: |
          modal run cicd.multigpu
--- a/.github/workflows/nightlies.yml
+++ b/.github/workflows/nightlies.yml
@@ -12,15 +12,26 @@ jobs:
      fail-fast: false
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.6.0
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.7.1
+            pytorch: 2.3.1
+            axolotl_extras:
+            is_latest: true
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.4.1
+            axolotl_extras:
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.5.1
            axolotl_extras:
    runs-on: axolotl-gpu-runner
    steps:
@@ -65,15 +76,26 @@ jobs:
    strategy:
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.6.0
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
            python_version: "3.11"
-            pytorch: 2.7.1
+            pytorch: 2.3.1
+            axolotl_extras:
+            is_latest: true
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.4.1
+            axolotl_extras:
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.5.1
            axolotl_extras:
    runs-on: axolotl-gpu-runner
    steps:
--- a/.github/workflows/precommit-autoupdate.yml
+++ b/.github/workflows/precommit-autoupdate.yml
@@ -1,40 +0,0 @@
-name: Pre-commit auto-update
-
-on:
-  schedule:
-    - cron: '0 0 * * 0'  # Run weekly
-  workflow_dispatch:  # Manual kickoff
-
-jobs:
-  auto-update:
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-      pull-requests: write
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/setup-python@v5
-        with:
-          python-version: '3.11'
-
-      - name: Update pre-commit hooks
-        id: update
-        run: |
-          pip install pre-commit
-          pre-commit autoupdate
-          if [[ -n $(git status --porcelain) ]]; then
-            echo "changes=true" >> $GITHUB_OUTPUT
-          fi
-
-      - name: Create Pull Request
-        if: steps.update.outputs.changes == 'true'
-        uses: peter-evans/create-pull-request@v6
-        with:
-          token: ${{ secrets.GITHUB_TOKEN }}
-          branch: update/pre-commit-hooks
-          delete-branch: true
-          title: "chore: update pre-commit hooks"
-          commit-message: "chore: update pre-commit hooks"
-          body: |
-            Automated PR to update pre-commit hooks to their latest versions.
--- a/.github/workflows/preview-docs.yml
+++ b/.github/workflows/preview-docs.yml
@@ -1,78 +0,0 @@
-name: Preview
-on:
-  workflow_dispatch:
-  pull_request:
-    types: [opened, synchronize, reopened, ready_for_review]
-
-    # Run the workflow only when one of these files changes
-    paths:
-      - '**/*.md'      # any Markdown file
-      - '**/*.qmd'     # any Quarto file
-      - '_quarto.yml'
-      - docs/scripts/generate_config_docs.py
-      - src/axolotl/utils/schemas/**.py
-
-permissions:
-  checks: write
-  contents: write
-  deployments: write
-  issues: write
-  discussions: write
-  pages: write
-  pull-requests: write
-  statuses: write
-
-jobs:
-  preview:
-    runs-on: ubuntu-latest
-    if: ${{ !github.event.pull_request.draft }}
-    steps:
-      - name: Check out repository
-        uses: actions/checkout@v4
-        with:
-          ref: ${{ github.event.pull_request.head.sha }}
-
-      - name: Set up Quarto
-        uses: quarto-dev/quarto-actions/setup@v2
-
-      - name: Setup Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: '3.11'
-
-      - name: Install dependencies
-        run: |
-          python3 -m pip install jupyter quartodoc
-          python3 -m pip install -e .
-
-      - name: Build autodoc
-        run: quartodoc build
-
-      - name: Quarto render
-        run: quarto render
-
-      - name: Netlify Publish
-        uses: nwtgck/actions-netlify@v3.0
-        if: ${{ github.event.pull_request.head.repo.full_name == github.repository }}
-        id: netlify
-        with:
-          publish-dir: './_site'
-          enable-pull-request-comment: false
-          enable-github-deployment: false
-          github-token: ${{ secrets.GITHUB_TOKEN }}
-          deploy-message: "Deployed On Netlify"
-          github-deployment-environment: 'preview'
-          github-deployment-description: 'Preview Deployment'
-        env:
-          NETLIFY_AUTH_TOKEN: ${{ secrets.NETLIFY_AUTH_TOKEN }}
-          NETLIFY_SITE_ID: ${{ secrets.NETLIFY_SITE_ID }}
-
-      - name: Update PR with preview link
-        if: ${{ steps.netlify.outcome == 'success' }}
-        uses: marocchino/sticky-pull-request-comment@v2
-        with:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          message: |
-            📖 **Documentation Preview**: ${{ steps.netlify.outputs.deploy-url }}
-
-            Deployed on Netlify from commit ${{ github.event.pull_request.head.sha }}
--- a/.github/workflows/pypi.yml
+++ b/.github/workflows/pypi.yml
@@ -36,11 +36,11 @@ jobs:
      - name: Setup Python
        uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"

      - name: Install dependencies
        run: |
-          pip3 install wheel packaging==23.2
+          pip3 install wheel packaging
          pip3 install --no-build-isolation -e .
          pip3 install -r requirements-dev.txt -r requirements-tests.txt

--- a/.github/workflows/tests-nightly.yml
+++ b/.github/workflows/tests-nightly.yml
@@ -12,7 +12,7 @@ jobs:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
          cache: 'pip' # caching pip dependencies
      - uses: pre-commit/action@v3.0.1
        env:
@@ -25,20 +25,19 @@ jobs:
      fail-fast: false
      max-parallel: 2
      matrix:
-        python_version: ["3.11"]
-        pytorch_version: ["2.6.0", "2.7.0"]
+        python_version: ["3.10", "3.11"]
+        pytorch_version: ["2.3.1", "2.4.1", "2.5.1"]
+        exclude:
+          - python_version: "3.10"
+            pytorch_version: "2.4.1"
+          - python_version: "3.10"
+            pytorch_version: "2.5.1"
    timeout-minutes: 20

    steps:
      - name: Check out repository code
        uses: actions/checkout@v4

-      - name: Restore Cache from S3
-        id: hf-cache-restore-s3
-        run: |
-          mkdir -p /home/runner/.cache/huggingface/hub
-          curl -L https://d1dttdx32dkk5p.cloudfront.net/hf-cache.tar.zst | tar -xf - -C /home/runner/.cache/huggingface/hub/  --use-compress-program unzstd
-
      - name: Setup Python
        uses: actions/setup-python@v5
        with:
@@ -48,11 +47,11 @@ jobs:
      - name: upgrade pip
        run: |
          pip3 install --upgrade pip
-          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 wheel
+          pip3 install --upgrade packaging setuptools wheel

      - name: Install PyTorch
        run: |
-          pip3 install torch==${{ matrix.pytorch_version }} torchvision
+          pip3 install torch==${{ matrix.pytorch_version }} --index-url https://download.pytorch.org/whl/cpu

      - name: Update requirements.txt
        run: |
@@ -64,7 +63,8 @@ jobs:

      - name: Install dependencies
        run: |
-          pip3 show torch
+          pip3 install --upgrade pip
+          pip3 install --upgrade packaging
          pip3 install --no-build-isolation -U -e .
          python scripts/unsloth_install.py | sh
          python scripts/cutcrossentropy_install.py | sh
@@ -80,9 +80,8 @@ jobs:

      - name: Run tests
        run: |
-          pytest -v --durations=10 -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ --ignore=tests/cli/ tests/
-          pytest -v --durations=10 tests/patched/
-          pytest -v --durations=10 tests/cli/
+          pytest -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ tests/
+          pytest tests/patched/

      - name: cleanup pip cache
        run: |
@@ -92,24 +91,31 @@ jobs:
    if: github.repository_owner == 'axolotl-ai-cloud'
    # this job needs to be run on self-hosted GPU runners...
    runs-on: [self-hosted, modal]
-    timeout-minutes: 120
+    timeout-minutes: 60
    needs: [pre-commit, pytest]

    strategy:
      fail-fast: false
      matrix:
        include:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
+            num_gpus: 1
+            axolotl_extras: mamba-ssm
+            nightly_build: "true"
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.4.1
            num_gpus: 1
            axolotl_extras:
            nightly_build: "true"
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 124
+            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.7.1
+            pytorch: 2.5.1
            num_gpus: 1
            axolotl_extras:
            nightly_build: "true"
@@ -119,11 +125,11 @@ jobs:
      - name: Install Python
        uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
      - name: Install Modal
        run: |
          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
+          pip install modal==0.71.8 jinja2
      - name: Update env vars
        run: |
          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
@@ -133,49 +139,6 @@ jobs:
          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
          echo "NIGHTLY_BUILD=${{ matrix.nightly_build }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
      - name: Run tests job on Modal
        run: |
-          modal run cicd.e2e_tests
-  docker-e2e-multigpu-tests:
-    if: github.repository_owner == 'axolotl-ai-cloud'
-    # this job needs to be run on self-hosted GPU runners...
-    runs-on: [self-hosted, modal]
-    timeout-minutes: 120
-    needs: [pre-commit, pytest, docker-e2e-tests]
-
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            num_gpus: 2
-            axolotl_extras:
-            nightly_build: "true"
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Install Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: "3.11"
-      - name: Install Modal
-        run: |
-          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
-      - name: Update env vars
-        run: |
-          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
-          echo "PYTORCH_VERSION=${{ matrix.pytorch}}" >> $GITHUB_ENV
-          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
-          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
-          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
-          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
-          echo "NIGHTLY_BUILD=${{ matrix.nightly_build }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
-      - name: Run tests job on Modal
-        run: |
-          modal run cicd.multigpu
+          modal run cicd.tests
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -13,7 +13,6 @@ on:
      - 'cicd/cicd.sh'
      - 'cicd/Dockerfile.jinja'
  pull_request:
-      types: [opened, synchronize, reopened, ready_for_review]
      paths:
       - '**.py'
       - 'requirements.txt'
@@ -28,19 +27,15 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}

-env:
-  TRANSFORMERS_IS_CI: "yes"
-
 jobs:
  pre-commit:
    name: pre-commit
    runs-on: ubuntu-latest
-    if: ${{ !github.event.pull_request.draft }}
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
          cache: 'pip' # caching pip dependencies
      - uses: pre-commit/action@v3.0.1
        env:
@@ -49,24 +44,31 @@ jobs:
  pytest:
    name: PyTest
    runs-on: ubuntu-latest
-    if: ${{ !github.event.pull_request.draft }}
-#    needs: [preload-cache]
    strategy:
      fail-fast: false
+      max-parallel: 2
      matrix:
-        python_version: ["3.11"]
-        pytorch_version: ["2.6.0", "2.7.0", "2.7.1"]
+        python_version: ["3.10", "3.11"]
+        pytorch_version: ["2.3.1", "2.4.1", "2.5.1"]
+        exclude:
+          - python_version: "3.10"
+            pytorch_version: "2.4.1"
+          - python_version: "3.10"
+            pytorch_version: "2.5.1"
    timeout-minutes: 20

    steps:
      - name: Check out repository code
        uses: actions/checkout@v4

-      - name: Restore Cache from S3
-        id: hf-cache-restore-s3
-        run: |
-          mkdir -p /home/runner/.cache/huggingface/hub
-          curl -L https://d1dttdx32dkk5p.cloudfront.net/hf-cache.tar.zst | tar -xf - -C /home/runner/.cache/huggingface/hub/  --use-compress-program unzstd
+      - name: Restore HF cache
+        id: hf-cache-restore
+        uses: actions/cache/restore@v4
+        with:
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ runner.os }}-hf-hub-cache-${{ hashFiles('**/conftest.py') }}

      - name: Setup Python
        uses: actions/setup-python@v5
@@ -77,11 +79,11 @@ jobs:
      - name: upgrade pip
        run: |
          pip3 install --upgrade pip
-          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 wheel
+          pip3 install --upgrade packaging setuptools wheel

      - name: Install PyTorch
        run: |
-          pip3 install torch==${{ matrix.pytorch_version }} torchvision
+          pip3 install torch==${{ matrix.pytorch_version }}

      - name: Install dependencies
        run: |
@@ -99,49 +101,47 @@ jobs:
        run: |
          axolotl --help

-      - name: Pre-Download dataset fixture
-        run: |
-          huggingface-cli download --repo-type=dataset axolotl-ai-internal/axolotl-oss-dataset-fixtures
-
      - name: Run tests
        run: |
-          pytest -v --durations=10 -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ --ignore=tests/cli/ --ignore=tests/monkeypatch/ tests/ --cov=axolotl --cov-report=xml
-          pytest -v --durations=10 tests/monkeypatch/ --cov=axolotl --cov-append --cov-report=xml
-          pytest -v --durations=10 tests/patched/ --cov=axolotl --cov-append --cov-report=xml
-          pytest -v --durations=10 tests/cli/ --cov=axolotl --cov-append --cov-report=xml
-
-      - name: Upload coverage to Codecov
-        uses: codecov/codecov-action@v5
-        with:
-          token: ${{ secrets.CODECOV_TOKEN }}
-          files: ./coverage.xml
-          flags: unittests,pytorch-${{ matrix.pytorch_version }}
-          fail_ci_if_error: false
+          pytest -v -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ tests/
+          pytest -v tests/patched/

      - name: cleanup pip cache
        run: |
          find "$(pip cache dir)/http-v2" -type f -mtime +14 -exec rm {} \;

+      - name: Save HF cache
+        id: hf-cache
+        uses: actions/cache/save@v4
+        with:
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ steps.hf-cache-restore.outputs.cache-primary-key }}
+
  pytest-sdist:
    name: PyTest from Source Dist
    runs-on: ubuntu-latest
-    if: ${{ !github.event.pull_request.draft }}
    strategy:
      fail-fast: false
+      max-parallel: 1
      matrix:
        python_version: ["3.11"]
-        pytorch_version: ["2.6.0", "2.7.0", "2.7.1"]
+        pytorch_version: ["2.4.1", "2.5.1"]
    timeout-minutes: 20

    steps:
      - name: Check out repository code
        uses: actions/checkout@v4

-      - name: Restore Cache from S3
-        id: hf-cache-restore-s3
-        run: |
-          mkdir -p /home/runner/.cache/huggingface/hub
-          curl -L https://d1dttdx32dkk5p.cloudfront.net/hf-cache.tar.zst | tar -xf - -C /home/runner/.cache/huggingface/hub/  --use-compress-program unzstd
+      - name: Restore HF cache
+        id: hf-cache-restore
+        uses: actions/cache/restore@v4
+        with:
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ runner.os }}-hf-hub-cache-${{ hashFiles('**/conftest.py') }}

      - name: Setup Python
        uses: actions/setup-python@v5
@@ -152,11 +152,11 @@ jobs:
      - name: upgrade pip
        run: |
          pip3 install --upgrade pip
-          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 setuptools_scm build wheel
+          pip3 install --upgrade packaging setuptools setuptools_scm build wheel

      - name: Install PyTorch
        run: |
-          pip3 install torch==${{ matrix.pytorch_version }} torchvision
+          pip3 install torch==${{ matrix.pytorch_version }}

      - name: Install dependencies
        run: |
@@ -175,160 +175,30 @@ jobs:
        run: |
          axolotl --help

-      - name: Show HF cache
-        run: huggingface-cli scan-cache
-
      - name: Run tests
        run: |
-          pytest -v --durations=10 -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ --ignore=tests/cli/ --ignore=tests/monkeypatch/ tests/ --cov=axolotl --cov-report=xml
-          pytest -v --durations=10 tests/monkeypatch/ --cov=axolotl --cov-append --cov-report=xml
-          pytest -v --durations=10 tests/cli/
+          pytest -v -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ tests/
+          pytest -v tests/patched/

      - name: cleanup pip cache
        run: |
          find "$(pip cache dir)/http-v2" -type f -mtime +14 -exec rm {} \;

-  gate-skip-e2e:
-    needs: [pre-commit, pytest, pytest-sdist]
-    runs-on: ubuntu-latest
-    outputs:
-      skip: ${{ steps.compute.outputs.skip }}
-    steps:
-      - uses: actions/github-script@v7
-        id: compute
+      - name: Save HF cache
+        id: hf-cache
+        uses: actions/cache/save@v4
        with:
-          script: |
-            const token = /\[skip-e2e\]/i;
-            let msg = '';
-            if (context.eventName === 'push') {
-              msg = context.payload.head_commit?.message || '';
-            } else if (context.eventName === 'pull_request') {
-              const { owner, repo } = context.repo;
-              const prNumber = context.payload.pull_request.number;
-              const commits = await github.paginate(
-                github.rest.pulls.listCommits,
-                { owner, repo, pull_number: prNumber, per_page: 100 }
-              );
-              msg = commits.at(-1)?.commit?.message || '';
-            }
-            const title = context.payload.pull_request?.title || '';
-            const body  = context.payload.pull_request?.body  || '';
-            const skip = token.test(msg) || token.test(title) || token.test(body);
-            core.setOutput('skip', String(skip));
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ steps.hf-cache-restore.outputs.cache-primary-key }}

  docker-e2e-tests-1st:
-    # Run this job first as a gate for running the remainder of the test matrix
-    if: >
-      github.repository_owner == 'axolotl-ai-cloud' &&
-      (github.event_name != 'pull_request' || !github.event.pull_request.draft) &&
-      needs.gate-skip-e2e.outputs.skip != 'true'
+    if: ${{ ! contains(github.event.commits[0].message, '[skip e2e]') && github.repository_owner == 'axolotl-ai-cloud' }}
    # this job needs to be run on self-hosted GPU runners...
-    runs-on: [self-hosted, modal]
-    timeout-minutes: 120
-    needs: [pre-commit, pytest, pytest-sdist, gate-skip-e2e]
-
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.7.1
-            num_gpus: 1
-            axolotl_extras:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.6.0
-            num_gpus: 1
-            axolotl_extras:
-            dockerfile: "Dockerfile-uv.jinja"
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Install Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: "3.11"
-      - name: Install Modal
-        run: |
-          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
-      - name: Update env vars
-        run: |
-          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
-          echo "PYTORCH_VERSION=${{ matrix.pytorch}}" >> $GITHUB_ENV
-          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
-          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
-          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
-          echo "MODAL_IMAGE_BUILDER_VERSION=2024.10" >> $GITHUB_ENV
-          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
-          echo "E2E_DOCKERFILE=${{ matrix.dockerfile || 'Dockerfile.jinja'}}" >> $GITHUB_ENV
-      - name: Run tests job on Modal
-        run: |
-          modal run cicd.e2e_tests
-
-  docker-e2e-tests:
-    if: >
-      github.repository_owner == 'axolotl-ai-cloud' &&
-      (github.event_name != 'pull_request' || !github.event.pull_request.draft) &&
-      needs.gate-skip-e2e.outputs.skip != 'true'
-    # this job needs to be run on self-hosted GPU runners...
-    runs-on: [self-hosted, modal]
-    timeout-minutes: 120
-    # Only run the remainder of the matrix if the first e2e check passed;
-    # this is to save on wasted compute costs for known failures that get caught in the first run
-    needs: [pre-commit, pytest, gate-skip-e2e, docker-e2e-tests-1st]
-
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - cuda: 126
-            cuda_version: 12.6.3
-            python_version: "3.11"
-            pytorch: 2.6.0
-            num_gpus: 1
-            axolotl_extras:
-          - cuda: 128
-            cuda_version: 12.8.1
-            python_version: "3.11"
-            pytorch: 2.7.1
-            num_gpus: 1
-            axolotl_extras:
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Install Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: "3.11"
-      - name: Install Modal
-        run: |
-          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
-      - name: Update env vars
-        run: |
-          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
-          echo "PYTORCH_VERSION=${{ matrix.pytorch}}" >> $GITHUB_ENV
-          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
-          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
-          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
-          echo "MODAL_IMAGE_BUILDER_VERSION=2024.10" >> $GITHUB_ENV
-          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
-          echo "E2E_DOCKERFILE=${{ matrix.dockerfile || 'Dockerfile.jinja'}}" >> $GITHUB_ENV
-      - name: Run tests job on Modal
-        run: |
-          modal run cicd.e2e_tests
-
-  docker-e2e-cleanup:
    runs-on: [self-hosted, modal]
    timeout-minutes: 90
-    needs: [docker-e2e-tests]
-    if: ${{ !github.event.pull_request.draft }}
+    needs: [pre-commit, pytest, pytest-sdist]

    strategy:
      fail-fast: false
@@ -337,7 +207,7 @@ jobs:
          - cuda: 124
            cuda_version: 12.4.1
            python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.4.1
            num_gpus: 1
            axolotl_extras:
    steps:
@@ -346,11 +216,11 @@ jobs:
      - name: Install Python
        uses: actions/setup-python@v5
        with:
-          python-version: "3.11"
+          python-version: "3.10"
      - name: Install Modal
        run: |
          python -m pip install --upgrade pip
-          pip install modal==1.0.2 jinja2
+          pip install modal==0.71.8 jinja2
      - name: Update env vars
        run: |
          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
@@ -358,9 +228,53 @@ jobs:
          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
-          echo "MODAL_IMAGE_BUILDER_VERSION=2024.10" >> $GITHUB_ENV
          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
-          echo "CODECOV_TOKEN=${{ secrets.CODECOV_TOKEN }}" >> $GITHUB_ENV
      - name: Run tests job on Modal
        run: |
-          modal run cicd.cleanup
+          modal run cicd.tests
+
+  docker-e2e-tests:
+    if: github.repository_owner == 'axolotl-ai-cloud'
+    # this job needs to be run on self-hosted GPU runners...
+    runs-on: [self-hosted, modal]
+    timeout-minutes: 90
+    needs: [pre-commit, pytest, docker-e2e-tests-1st]
+
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - cuda: 121
+            cuda_version: 12.1.1
+            python_version: "3.10"
+            pytorch: 2.3.1
+            num_gpus: 1
+            axolotl_extras: mamba-ssm
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.5.1
+            num_gpus: 1
+            axolotl_extras:
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Install Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.10"
+      - name: Install Modal
+        run: |
+          python -m pip install --upgrade pip
+          pip install modal==0.71.8 jinja2
+      - name: Update env vars
+        run: |
+          echo "BASE_TAG=main-base-py${{ matrix.python_version }}-cu${{ matrix.cuda }}-${{ matrix.pytorch }}" >> $GITHUB_ENV
+          echo "PYTORCH_VERSION=${{ matrix.pytorch}}" >> $GITHUB_ENV
+          echo "AXOLOTL_ARGS=${{ matrix.axolotl_args}}" >> $GITHUB_ENV
+          echo "AXOLOTL_EXTRAS=${{ matrix.axolotl_extras}}" >> $GITHUB_ENV
+          echo "CUDA=${{ matrix.cuda }}" >> $GITHUB_ENV
+          echo "N_GPUS=${{ matrix.num_gpus }}" >> $GITHUB_ENV
+      - name: Run tests job on Modal
+        run: |
+          modal run cicd.tests
--- a/.gitignore
+++ b/.gitignore
@@ -181,12 +181,11 @@ prepared-datasets/
 submit.sh
 *.out*

-# Quartodoc generated files
-objects.json
-site_libs/
-
 typings/
 out/

 # vim
 *.swp
+
+# symlinked to axolotl-artifacts in docker containers
+outputs
--- a/.isort.cfg
+++ b/.isort.cfg
@@ -1,4 +1,3 @@
 [settings]
 profile=black
 known_third_party=wandb,comet_ml
-known_local_folder=src,tests
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -3,7 +3,7 @@ default_language_version:

 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v6.0.0
+    rev: v4.4.0
    hooks:
    -   id: check-yaml
    -   id: end-of-file-fixer
@@ -11,23 +11,23 @@ repos:
    -   id: no-commit-to-branch
        args: ['--branch', 'main']
 -   repo: https://github.com/psf/black
-    rev: 25.1.0
+    rev: 23.3.0
    hooks:
    -   id: black
 -   repo: https://github.com/pycqa/isort
-    rev: 6.0.1
+    rev: 5.12.0
    hooks:
      - id: isort
 -   repo: https://github.com/PyCQA/flake8
-    rev: 7.3.0
+    rev: 6.0.0
    hooks:
    - id: flake8
-   repo: https://github.com/pylint-dev/pylint
-    rev: v3.3.8
+-   repo: https://github.com/PyCQA/pylint
+    rev: v3.3.0
    hooks:
    - id: pylint
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.17.1
+    rev: v1.3.0
    hooks:
    - id: mypy
      additional_dependencies:
@@ -36,7 +36,7 @@ repos:
            'pydantic>=2.5.3',
        ]
 -   repo: https://github.com/PyCQA/bandit
-    rev: 1.8.6
+    rev: 1.7.5
    hooks:
    -   id: bandit
        args: [
--- a/.runpod/.gitignore
+++ b/.runpod/.gitignore
@@ -1,161 +0,0 @@
-# Byte-compiled / optimized / DLL files
-__pycache__/
-*.py[cod]
-*$py.class
-
-# C extensions
-*.so
-
-# Distribution / packaging
-.Python
-build/
-develop-eggs/
-dist/
-downloads/
-eggs/
-.eggs/
-lib/
-lib64/
-parts/
-sdist/
-var/
-wheels/
-share/python-wheels/
-*.egg-info/
-.installed.cfg
-*.egg
-MANIFEST
-
-# PyInstaller
-#  Usually these files are written by a python script from a template
-#  before PyInstaller builds the exe, so as to inject date/other infos into it.
-*.manifest
-*.spec
-
-# Installer logs
-pip-log.txt
-pip-delete-this-directory.txt
-
-# Unit test / coverage reports
-htmlcov/
-.tox/
-.nox/
-.coverage
-.coverage.*
-.cache
-nosetests.xml
-coverage.xml
-*.cover
-*.py,cover
-.hypothesis/
-.pytest_cache/
-cover/
-
-# Translations
-*.mo
-*.pot
-
-# Django stuff:
-*.log
-local_settings.py
-db.sqlite3
-db.sqlite3-journal
-
-# Flask stuff:
-instance/
-.webassets-cache
-
-# Scrapy stuff:
-.scrapy
-
-# Sphinx documentation
-docs/_build/
-
-# PyBuilder
-.pybuilder/
-target/
-
-# Jupyter Notebook
-.ipynb_checkpoints
-
-# IPython
-profile_default/
-ipython_config.py
-
-# pyenv
-#   For a library or package, you might want to ignore these files since the code is
-#   intended to run in multiple environments; otherwise, check them in:
-# .python-version
-
-# pipenv
-#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
-#   However, in case of collaboration, if having platform-specific dependencies or dependencies
-#   having no cross-platform support, pipenv may install dependencies that don't work, or not
-#   install all needed dependencies.
-#Pipfile.lock
-
-# poetry
-#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
-#   This is especially recommended for binary packages to ensure reproducibility, and is more
-#   commonly ignored for libraries.
-#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
-#poetry.lock
-
-# pdm
-#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
-#pdm.lock
-#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
-#   in version control.
-#   https://pdm.fming.dev/#use-with-ide
-.pdm.toml
-
-# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
-__pypackages__/
-
-# Celery stuff
-celerybeat-schedule
-celerybeat.pid
-
-# SageMath parsed files
-*.sage.py
-
-# Environments
-.env
-.venv
-env/
-venv/
-ENV/
-env.bak/
-venv.bak/
-
-# Spyder project settings
-.spyderproject
-.spyproject
-
-# Rope project settings
-.ropeproject
-
-# mkdocs documentation
-/site
-
-# mypy
-.mypy_cache/
-.dmypy.json
-dmypy.json
-
-# Pyre type checker
-.pyre/
-
-# pytype static type analyzer
-.pytype/
-
-# Cython debug symbols
-cython_debug/
-
-# PyCharm
-#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
-#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
-#  and can be added to the global gitignore or merged into this file.  For a more nuclear
-#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
-#.idea/
-pod/scripts/config.yaml
--- a/.runpod/Dockerfile
+++ b/.runpod/Dockerfile
@@ -1,18 +0,0 @@
-FROM axolotlai/axolotl-cloud:main-py3.11-cu124-2.6.0
-
-COPY .runpod/requirements.txt /requirements.txt
-RUN --mount=type=cache,target=/root/.cache/pip \
-    python3 -m pip install --upgrade pip && \
-    python3 -m pip install --upgrade -r /requirements.txt
-
-# Environment settings
-ARG BASE_VOLUME="/runpod-volume"
-ENV BASE_VOLUME=$BASE_VOLUME
-ENV HF_DATASETS_CACHE="${BASE_VOLUME}/huggingface-cache/datasets"
-ENV HUGGINGFACE_HUB_CACHE="${BASE_VOLUME}/huggingface-cache/hub"
-ENV TRANSFORMERS_CACHE="${BASE_VOLUME}/huggingface-cache/hub"
-
-COPY .runpod/src /src
-
-WORKDIR /src
-CMD ["python3", "/src/handler.py"]
--- a/.runpod/README.md
+++ b/.runpod/README.md
@@ -1,335 +0,0 @@
-<h1>LLM Post Training- Full fine-tune, LoRA, QLoRa etc. Llama/Mistral/Gemma and more</h1>
-
-# Configuration Options
-
-This document outlines all available configuration options for training models. The configuration can be provided as a JSON request.
-
-## Usage
-
-You can use these configuration Options:
-
-1. As a JSON request body:
-
-```json
-{
-  "input": {
-    "user_id": "user",
-    "model_id": "model-name",
-    "run_id": "run-id",
-    "credentials": {
-      "wandb_api_key": "", # add your Weights & biases key. TODO:  you will be able to set this in Enviornment variables.
-      "hf_token": "", # add your HF_token. TODO:  you will be able to set this in Enviornment variables.
-    },
-    "args": {
-      "base_model": "NousResearch/Llama-3.2-1B",
-      // ... other options
-    }
-  }
-}
-```
-
-## Configuration Options
-
-### Model Configuration
-
-| Option              | Description                                                                                   | Default              |
-| ------------------- | --------------------------------------------------------------------------------------------- | -------------------- |
-| `base_model`        | Path to the base model (local or HuggingFace)                                                 | Required             |
-| `base_model_config` | Configuration path for the base model                                                         | Same as base_model   |
-| `revision_of_model` | Specific model revision from HuggingFace hub                                                  | Latest               |
-| `tokenizer_config`  | Custom tokenizer configuration path                                                           | Optional             |
-| `model_type`        | Type of model to load                                                                         | AutoModelForCausalLM |
-| `tokenizer_type`    | Type of tokenizer to use                                                                      | AutoTokenizer        |
-| `hub_model_id`      | Repository ID where the model will be pushed on Hugging Face Hub (format: username/repo-name) | Optional             |
-
-## Model Family Identification
-
-| Option                     | Default | Description                    |
-| -------------------------- | ------- | ------------------------------ |
-| `is_falcon_derived_model`  | `false` | Whether model is Falcon-based  |
-| `is_llama_derived_model`   | `false` | Whether model is LLaMA-based   |
-| `is_qwen_derived_model`    | `false` | Whether model is Qwen-based    |
-| `is_mistral_derived_model` | `false` | Whether model is Mistral-based |
-
-## Model Configuration Overrides
-
-| Option                                          | Default    | Description                        |
-| ----------------------------------------------- | ---------- | ---------------------------------- |
-| `overrides_of_model_config.rope_scaling.type`   | `"linear"` | RoPE scaling type (linear/dynamic) |
-| `overrides_of_model_config.rope_scaling.factor` | `1.0`      | RoPE scaling factor                |
-
-### Model Loading Options
-
-| Option         | Description                   | Default |
-| -------------- | ----------------------------- | ------- |
-| `load_in_8bit` | Load model in 8-bit precision | false   |
-| `load_in_4bit` | Load model in 4-bit precision | false   |
-| `bf16`         | Use bfloat16 precision        | false   |
-| `fp16`         | Use float16 precision         | false   |
-| `tf32`         | Use tensor float 32 precision | false   |
-
-## Memory and Device Settings
-
-| Option             | Default   | Description             |
-| ------------------ | --------- | ----------------------- |
-| `gpu_memory_limit` | `"20GiB"` | GPU memory limit        |
-| `lora_on_cpu`      | `false`   | Load LoRA on CPU        |
-| `device_map`       | `"auto"`  | Device mapping strategy |
-| `max_memory`       | `null`    | Max memory per device   |
-
-## Training Hyperparameters
-
-| Option                        | Default   | Description                 |
-| ----------------------------- | --------- | --------------------------- |
-| `gradient_accumulation_steps` | `1`       | Gradient accumulation steps |
-| `micro_batch_size`            | `2`       | Batch size per GPU          |
-| `eval_batch_size`             | `null`    | Evaluation batch size       |
-| `num_epochs`                  | `4`       | Number of training epochs   |
-| `warmup_steps`                | `100`     | Warmup steps                |
-| `warmup_ratio`                | `0.05`    | Warmup ratio                |
-| `learning_rate`               | `0.00003` | Learning rate               |
-| `lr_quadratic_warmup`         | `false`   | Quadratic warmup            |
-| `logging_steps`               | `null`    | Logging frequency           |
-| `eval_steps`                  | `null`    | Evaluation frequency        |
-| `evals_per_epoch`             | `null`    | Evaluations per epoch       |
-| `save_strategy`               | `"epoch"` | Checkpoint saving strategy  |
-| `save_steps`                  | `null`    | Saving frequency            |
-| `saves_per_epoch`             | `null`    | Saves per epoch             |
-| `save_total_limit`            | `null`    | Maximum checkpoints to keep |
-| `max_steps`                   | `null`    | Maximum training steps      |
-
-### Dataset Configuration
-
-```yaml
-datasets:
-  - path: vicgalle/alpaca-gpt4 # HuggingFace dataset or TODO: You will be able to add the local path.
-    type: alpaca # Format type (alpaca, gpteacher, oasst, etc.)
-    ds_type: json # Dataset type
-    data_files: path/to/data # Source data files
-    train_on_split: train # Dataset split to use
-```
-
-## Chat Template Settings
-
-| Option                   | Default                          | Description            |
-| ------------------------ | -------------------------------- | ---------------------- |
-| `chat_template`          | `"tokenizer_default"`            | Chat template type     |
-| `chat_template_jinja`    | `null`                           | Custom Jinja template  |
-| `default_system_message` | `"You are a helpful assistant."` | Default system message |
-
-## Dataset Processing
-
-| Option                            | Default                    | Description                         |
-| --------------------------------- | -------------------------- | ----------------------------------- |
-| `dataset_prepared_path`           | `"data/last_run_prepared"` | Path for prepared dataset           |
-| `push_dataset_to_hub`             | `""`                       | Push dataset to HF hub              |
-| `dataset_processes`               | `4`                        | Number of preprocessing processes   |
-| `dataset_keep_in_memory`          | `false`                    | Keep dataset in memory              |
-| `shuffle_merged_datasets`         | `true`                     | Shuffle merged datasets             |
-| `shuffle_before_merging_datasets` | `false`                    | Shuffle each dataset before merging |
-| `dataset_exact_deduplication`     | `true`                     | Deduplicate datasets                |
-
-## LoRA Configuration
-
-| Option                     | Default                | Description                    |
-| -------------------------- | ---------------------- | ------------------------------ |
-| `adapter`                  | `"lora"`               | Adapter type (lora/qlora)      |
-| `lora_model_dir`           | `""`                   | Directory with pretrained LoRA |
-| `lora_r`                   | `8`                    | LoRA attention dimension       |
-| `lora_alpha`               | `16`                   | LoRA alpha parameter           |
-| `lora_dropout`             | `0.05`                 | LoRA dropout                   |
-| `lora_target_modules`      | `["q_proj", "v_proj"]` | Modules to apply LoRA          |
-| `lora_target_linear`       | `false`                | Target all linear modules      |
-| `peft_layers_to_transform` | `[]`                   | Layers to transform            |
-| `lora_modules_to_save`     | `[]`                   | Modules to save                |
-| `lora_fan_in_fan_out`      | `false`                | Fan in/out structure           |
-
-## Optimization Settings
-
-| Option                    | Default | Description                |
-| ------------------------- | ------- | -------------------------- |
-| `train_on_inputs`         | `false` | Train on input prompts     |
-| `group_by_length`         | `false` | Group by sequence length   |
-| `gradient_checkpointing`  | `false` | Use gradient checkpointing |
-| `early_stopping_patience` | `3`     | Early stopping patience    |
-
-## Learning Rate Scheduling
-
-| Option                     | Default    | Description          |
-| -------------------------- | ---------- | -------------------- |
-| `lr_scheduler`             | `"cosine"` | Scheduler type       |
-| `lr_scheduler_kwargs`      | `{}`       | Scheduler parameters |
-| `cosine_min_lr_ratio`      | `null`     | Minimum LR ratio     |
-| `cosine_constant_lr_ratio` | `null`     | Constant LR ratio    |
-| `lr_div_factor`            | `null`     | LR division factor   |
-
-## Optimizer Settings
-
-| Option                 | Default      | Description         |
-| ---------------------- | ------------ | ------------------- |
-| `optimizer`            | `"adamw_hf"` | Optimizer choice    |
-| `optim_args`           | `{}`         | Optimizer arguments |
-| `optim_target_modules` | `[]`         | Target modules      |
-| `weight_decay`         | `null`       | Weight decay        |
-| `adam_beta1`           | `null`       | Adam beta1          |
-| `adam_beta2`           | `null`       | Adam beta2          |
-| `adam_epsilon`         | `null`       | Adam epsilon        |
-| `max_grad_norm`        | `null`       | Gradient clipping   |
-
-## Attention Implementations
-
-| Option                     | Default | Description                   |
-| -------------------------- | ------- | ----------------------------- |
-| `flash_optimum`            | `false` | Use better transformers       |
-| `xformers_attention`       | `false` | Use xformers                  |
-| `flash_attention`          | `false` | Use flash attention           |
-| `flash_attn_cross_entropy` | `false` | Flash attention cross entropy |
-| `flash_attn_rms_norm`      | `false` | Flash attention RMS norm      |
-| `flash_attn_fuse_mlp`      | `false` | Fuse MLP operations           |
-| `sdp_attention`            | `false` | Use scaled dot product        |
-| `s2_attention`             | `false` | Use shifted sparse attention  |
-
-## Tokenizer Modifications
-
-| Option           | Default | Description                  |
-| ---------------- | ------- | ---------------------------- |
-| `special_tokens` | -       | Special tokens to add/modify |
-| `tokens`         | `[]`    | Additional tokens            |
-
-## Distributed Training
-
-| Option                  | Default | Description           |
-| ----------------------- | ------- | --------------------- |
-| `fsdp`                  | `null`  | FSDP configuration    |
-| `fsdp_config`           | `null`  | FSDP config options   |
-| `deepspeed`             | `null`  | Deepspeed config path |
-| `ddp_timeout`           | `null`  | DDP timeout           |
-| `ddp_bucket_cap_mb`     | `null`  | DDP bucket capacity   |
-| `ddp_broadcast_buffers` | `null`  | DDP broadcast buffers |
-
-<details>
-<summary><h3>Example Configuration Request:</h3></summary>
-
-Here's a complete example for fine-tuning a LLaMA model using LoRA:
-
-```json
-{
-  "input": {
-    "user_id": "user",
-    "model_id": "llama-test",
-    "run_id": "test-run",
-    "credentials": {
-      "wandb_api_key": "",
-      "hf_token": ""
-    },
-    "args": {
-      "base_model": "NousResearch/Llama-3.2-1B",
-      "load_in_8bit": false,
-      "load_in_4bit": false,
-      "strict": false,
-      "datasets": [
-        {
-          "path": "teknium/GPT4-LLM-Cleaned",
-          "type": "alpaca"
-        }
-      ],
-      "dataset_prepared_path": "last_run_prepared",
-      "val_set_size": 0.1,
-      "output_dir": "./outputs/lora-out",
-      "adapter": "lora",
-      "sequence_len": 2048,
-      "sample_packing": true,
-      "eval_sample_packing": true,
-      "pad_to_sequence_len": true,
-      "lora_r": 16,
-      "lora_alpha": 32,
-      "lora_dropout": 0.05,
-      "lora_target_modules": [
-        "gate_proj",
-        "down_proj",
-        "up_proj",
-        "q_proj",
-        "v_proj",
-        "k_proj",
-        "o_proj"
-      ],
-      "gradient_accumulation_steps": 2,
-      "micro_batch_size": 2,
-      "num_epochs": 1,
-      "optimizer": "adamw_8bit",
-      "lr_scheduler": "cosine",
-      "learning_rate": 0.0002,
-      "train_on_inputs": false,
-      "group_by_length": false,
-      "bf16": "auto",
-      "tf32": false,
-      "gradient_checkpointing": true,
-      "logging_steps": 1,
-      "flash_attention": true,
-      "loss_watchdog_threshold": 5,
-      "loss_watchdog_patience": 3,
-      "warmup_steps": 10,
-      "evals_per_epoch": 4,
-      "saves_per_epoch": 1,
-      "weight_decay": 0,
-      "hub_model_id": "runpod/llama-fr-lora",
-      "wandb_name": "test-run-1",
-      "wandb_project": "test-run-1",
-      "wandb_entity": "axo-test",
-      "special_tokens": {
-        "pad_token": "<|end_of_text|>"
-      }
-    }
-  }
-}
-```
-
-</details>
-
-### Advanced Features
-
-#### Wandb Integration
-
- `wandb_project`: Project name for Weights & Biases
- `wandb_entity`: Team name in W&B
- `wandb_watch`: Monitor model with W&B
- `wandb_name`: Name of the W&B run
- `wandb_run_id`: ID for the W&B run
-
-#### Performance Optimization
-
- `sample_packing`: Enable efficient sequence packing
- `eval_sample_packing`: Use sequence packing during evaluation
- `torch_compile`: Enable PyTorch 2.0 compilation
- `flash_attention`: Use Flash Attention implementation
- `xformers_attention`: Use xFormers attention implementation
-
-### Available Optimizers
-
-The following optimizers are supported:
-
- `adamw_hf`: HuggingFace's AdamW implementation
- `adamw_torch`: PyTorch's AdamW
- `adamw_torch_fused`: Fused AdamW implementation
- `adamw_torch_xla`: XLA-optimized AdamW
- `adamw_apex_fused`: NVIDIA Apex fused AdamW
- `adafactor`: Adafactor optimizer
- `adamw_anyprecision`: Anyprecision AdamW
- `adamw_bnb_8bit`: 8-bit AdamW from bitsandbytes
- `lion_8bit`: 8-bit Lion optimizer
- `lion_32bit`: 32-bit Lion optimizer
- `sgd`: Stochastic Gradient Descent
- `adagrad`: Adagrad optimizer
-
-## Notes
-
- Set `load_in_8bit: true` or `load_in_4bit: true` for memory-efficient training
- Enable `flash_attention: true` for faster training on modern GPUs
- Use `gradient_checkpointing: true` to reduce memory usage
- Adjust `micro_batch_size` and `gradient_accumulation_steps` based on your GPU memory
-
-For more detailed information, please refer to the [documentation](https://axolotl-ai-cloud.github.io/axolotl/docs/config-reference.html).
-
-### Errors:
-
- if you face any issues with the Flash Attention-2, Delete yoor worker and Re-start.
--- a/.runpod/hub.json
+++ b/.runpod/hub.json
@@ -1,93 +0,0 @@
-{
-  "title": "Axolotl Fine-Tuning",
-  "description": "Serverless fine-tuning of open-source LLMs with Axolotl. Supports LoRA, QLoRA, DPO, and more using Hugging Face models and datasets.",
-  "type": "serverless",
-  "category": "language",
-  "iconUrl": "https://avatars.githubusercontent.com/u/167502477",
-  "config": {
-    "runsOn": "GPU",
-    "containerDiskInGb": 200,
-    "gpuCount": 1,
-    "allowedCudaVersions": [
-      "12.8",
-      "12.7",
-      "12.6",
-      "12.5",
-      "12.4"
-    ],
-    "presets": [],
-    "env": [
-      {
-        "key": "TOKENIZER",
-        "input": {
-          "name": "Tokenizer",
-          "type": "string",
-          "description": "Name or path of the Hugging Face tokenizer to use.",
-          "default": "",
-          "advanced": true
-        }
-      },
-      {
-        "key": "MAX_NUM_SEQS",
-        "input": {
-          "name": "Max Num Seqs",
-          "type": "number",
-          "description": "Maximum number of sequences per iteration.",
-          "default": 256,
-          "advanced": true
-        }
-      },
-      {
-        "key": "DISABLE_LOG_STATS",
-        "input": {
-          "name": "Disable Log Stats",
-          "type": "boolean",
-          "description": "Disable logging statistics.",
-          "default": false,
-          "trueValue": "true",
-          "falseValue": "false"
-        }
-      },
-      {
-        "key": "LOAD_FORMAT",
-        "input": {
-          "name": "Load Format",
-          "type": "string",
-          "description": "The format of the model weights to load.",
-          "default": "auto",
-          "options": [
-            {
-              "label": "auto",
-              "value": "auto"
-            },
-            {
-              "label": "pt",
-              "value": "pt"
-            },
-            {
-              "label": "safetensors",
-              "value": "safetensors"
-            },
-            {
-              "label": "npcache",
-              "value": "npcache"
-            },
-            {
-              "label": "dummy",
-              "value": "dummy"
-            },
-            {
-              "label": "tensorizer",
-              "value": "tensorizer"
-            },
-            {
-              "label": "bitsandbytes",
-              "value": "bitsandbytes"
-            }
-          ],
-          "advanced": true
-        }
-      }
-    ]
-  }
-}
--- a/.runpod/requirements.txt
+++ b/.runpod/requirements.txt
@@ -1,7 +0,0 @@
-# Required Python packages get listed here, one per line.
-# Reccomended to lock the version number to avoid unexpected changes.
-
-# You can also install packages from a git repository, e.g.:
-# git+https://github.com/runpod/runpod-python.git
-# To learn more, see https://pip.pypa.io/en/stable/reference/requirements-file-format/
-runpod~=1.7.0
--- a/.runpod/src/config/config.yaml
+++ b/.runpod/src/config/config.yaml
@@ -1,571 +0,0 @@
-# # This is the huggingface model that contains *.pt, *.safetensors, or *.bin files
-# # This can also be a relative path to a model on disk
-# base_model: ./llama-7b-hf
-# # You can specify an ignore pattern if the model repo contains more than 1 model type (*.pt, etc)
-# base_model_ignore_patterns:
-# # If the base_model repo on hf hub doesn't include configuration .json files,
-# # You can set that here, or leave this empty to default to base_model
-# base_model_config: ./llama-7b-hf
-# # You can specify to choose a specific model revision from huggingface hub
-# model_revision:
-# # Optional tokenizer configuration override in case you want to use a different tokenizer
-# # than the one defined in the base model
-# tokenizer_config:
-# # If you want to specify the type of model to load, AutoModelForCausalLM is a good choice too
-# model_type: AutoModelForCausalLM
-# # Corresponding tokenizer for the model AutoTokenizer is a good choice
-# tokenizer_type: AutoTokenizer
-# # Trust remote code for untrusted source
-# trust_remote_code:
-# # use_fast option for tokenizer loading from_pretrained, default to True
-# tokenizer_use_fast:
-# # Whether to use the legacy tokenizer setting, defaults to True
-# tokenizer_legacy:
-# # Resize the model embeddings when new tokens are added to multiples of 32
-# # This is reported to improve training speed on some models
-# resize_token_embeddings_to_32x:
-
-# # Used to identify which the model is based on
-# is_falcon_derived_model:
-# is_llama_derived_model:
-# # Please note that if you set this to true, `padding_side` will be set to "left" by default
-# is_mistral_derived_model:
-# is_qwen_derived_model:
-
-# # optional overrides to the base model configuration
-# model_config:
-#   # RoPE Scaling https://github.com/huggingface/transformers/pull/24653
-#   rope_scaling:
-#     type: # linear | dynamic
-#     factor: # float
-
-
-# # Whether you are training a 4-bit GPTQ quantized model
-# gptq: true
-# gptq_groupsize: 128 # group size
-# gptq_model_v1: false # v1 or v2
-
-# # This will attempt to quantize the model down to 8 bits and use adam 8 bit optimizer
-# load_in_8bit: true
-# # Use bitsandbytes 4 bit
-# load_in_4bit:
-
-# # Use CUDA bf16
-# bf16: true # bool or 'full' for `bf16_full_eval`. require >=ampere
-# # Use CUDA fp16
-# fp16: true
-# # Use CUDA tf32
-# tf32: true # require >=ampere
-
-# # No AMP (automatic mixed precision)
-# bfloat16: true # require >=ampere
-# float16: true
-
-# # A list of one or more datasets to finetune the model with
-# datasets:
-#   # HuggingFace dataset repo | s3://,gs:// path | "json" for local dataset, make sure to fill data_files
-#   - path: vicgalle/alpaca-gpt4
-#   # The type of prompt to use for training. [alpaca, sharegpt, gpteacher, oasst, reflection]
-#     type: alpaca # format | format:<prompt_style> (chat/instruct) | <prompt_strategies>.load_<load_fn>
-#     ds_type: # Optional[str] (json|arrow|parquet|text|csv) defines the datatype when path is a file
-#     data_files: # Optional[str] path to source data files
-#     shards: # Optional[int] number of shards to split data into
-#     name: # Optional[str] name of dataset configuration to load
-#     train_on_split: train # Optional[str] name of dataset split to load from
-
-#     # Optional[str] fastchat conversation type, only used with type: sharegpt
-#     conversation:  # Options (see Conversation 'name'): https://github.com/lm-sys/FastChat/blob/main/fastchat/conversation.py
-#     field_human: # Optional[str]. Human key to use for conversation.
-#     field_model: # Optional[str]. Assistant key to use for conversation.
-
-#   # Custom user prompt
-#   - path: repo
-#     type:
-#       # The below are defaults. only set what's needed.
-#       system_prompt: ""
-#       system_format: "{system}"
-#       field_system: system
-#       field_instruction: instruction
-#       field_input: input
-#       field_output: output
-
-#       # Customizable to be single line or multi-line
-#       # 'format' can include {input}
-#       format: |-
-#         User: {instruction} {input}
-#         Assistant:
-#       # 'no_input_format' cannot include {input}
-#       no_input_format: "{instruction} "
-
-#       # For `completion` datasets only, uses the provided field instead of `text` column
-#       field:
-
-# # Axolotl attempts to save the dataset as an arrow after packing the data together so
-# # subsequent training attempts load faster, relative path
-# dataset_prepared_path: data/last_run_prepared
-# # Push prepared dataset to hub
-# push_dataset_to_hub: # repo path
-# # The maximum number of processes to use while preprocessing your input dataset. This defaults to `os.cpu_count()`
-# # if not set.
-# dataset_processes: # defaults to os.cpu_count() if not set
-# # push checkpoints to hub
-# hub_model_id: # repo path to push finetuned model
-# # how to push checkpoints to hub
-# # https://huggingface.co/docs/transformers/v4.31.0/en/main_classes/trainer#transformers.TrainingArguments.hub_strategy
-# hub_strategy:
-# # Whether to use hf `use_auth_token` for loading datasets. Useful for fetching private datasets
-# # Required to be true when used in combination with `push_dataset_to_hub`
-# hf_use_auth_token: # boolean
-# # How much of the dataset to set aside as evaluation. 1 = 100%, 0.50 = 50%, etc. 0 for no eval.
-# val_set_size: 0.04
-# # Num shards for whole dataset
-# dataset_shard_num:
-# # Index of shard to use for whole dataset
-# dataset_shard_idx:
-
-# # The maximum length of an input to train with, this should typically be less than 2048
-# # as most models have a token/context limit of 2048
-# sequence_len: 2048
-# # Pad inputs so each step uses constant sized buffers
-# # This will reduce memory fragmentation and may prevent OOMs, by re-using memory more efficiently
-# pad_to_sequence_len:
-# # Max sequence length to concatenate training samples together up to
-# # Inspired by StackLLaMA. see https://huggingface.co/blog/stackllama#supervised-fine-tuning
-# # FutureWarning: This will soon be DEPRECATED
-# max_packed_sequence_len: 1024
-# # Use efficient multi-packing with block diagonal attention and per sequence position_ids. Recommend set to 'true'
-# sample_packing:
-# # Set to 'false' if getting errors during eval with sample_packing on.
-# eval_sample_packing:
-# # You can set these packing optimizations AFTER starting a training at least once.
-# # The trainer will provide recommended values for these values.
-# sample_packing_eff_est:
-# total_num_tokens:
-
-# # If you want to use 'lora' or 'qlora' or leave blank to train all parameters in original model
-# adapter: lora
-# # If you already have a lora model trained that you want to load, put that here.
-# # This means after training, if you want to test the model, you should set this to the value of `lora_out_dir`.
-# lora_model_dir:
-
-# # LoRA hyperparameters
-# # For more details about the following options, see:
-# # https://www.anyscale.com/blog/fine-tuning-llms-lora-or-full-parameter-an-in-depth-analysis-with-llama-2
-# lora_r: 8
-# lora_alpha: 16
-# lora_dropout: 0.05
-# lora_target_modules:
-#   - q_proj
-#   - v_proj
-# #  - k_proj
-# #  - o_proj
-# #  - gate_proj
-# #  - down_proj
-# #  - up_proj
-# lora_target_linear: # If true, will target all linear layers
-
-# # If you added new tokens to the tokenizer, you may need to save some LoRA modules because they need to know the new tokens.
-# # For LLaMA and Mistral, you need to save `embed_tokens` and `lm_head`. It may vary for other models.
-# # `embed_tokens` converts tokens to embeddings, and `lm_head` converts embeddings to token probabilities.
-# # https://github.com/huggingface/peft/issues/334#issuecomment-1561727994
-# lora_modules_to_save:
-# #  - embed_tokens
-# #  - lm_head
-
-# # Once you complete training, the model will be saved to the following directory.
-# # If you merge the adapter to the base model, a subdirectory `merged` will be created under this directory.
-# # Make sure `lora_model_dir` points to this directory if you want to use the trained model.
-# lora_out_dir:
-# lora_fan_in_fan_out: false
-
-# # ReLoRA configuration
-# # Must use either 'lora' or 'qlora' adapter, and does not support fsdp or deepspeed
-# relora_steps: # Number of steps per ReLoRA restart
-# relora_warmup_steps: # Number of per-restart warmup steps
-# relora_cpu_offload: # True to perform lora weight merges on cpu during restarts, for modest gpu memory savings
-
-# # wandb configuration if you're using it
-# wandb_mode: # "offline" to save run metadata locally and not sync to the server, "disabled" to turn off wandb
-# wandb_project: # Your wandb project name
-# wandb_entity: # A wandb Team name if using a Team
-# wandb_watch:
-# wandb_run_id: # Set the name of your wandb run
-# wandb_log_model: # "checkpoint" to log model to wandb Artifacts every `save_steps` or "end" to log only at the end of training
-
-# # Where to save the full-finetuned model to
-# output_dir: ./completed-model
-
-# # Whether to use torch.compile and which backend to use
-# torch_compile:  # bool
-# torch_compile_backend:  # Optional[str]
-
-# # Training hyperparameters
-
-# # If greater than 1, backpropagation will be skipped and the gradients will be accumulated for the given number of steps.
-# gradient_accumulation_steps: 1
-# # The number of samples to include in each batch. This is the number of samples sent to each GPU.
-# micro_batch_size: 2
-# eval_batch_size:
-# num_epochs: 4
-# warmup_steps: 100  # cannot use with warmup_ratio
-# warmup_ratio: 0.05  # cannot use with warmup_steps
-# learning_rate: 0.00003
-# lr_quadratic_warmup:
-# logging_steps:
-# save_strategy: # Set to `no` to skip checkpoint saves
-# save_steps: # Leave empty to save at each epoch
-# eval_steps: # Leave empty to eval at each epoch, integers for every N steps. decimal for fraction of total steps
-# save_total_limit: # Checkpoints saved at a time
-# # Maximum number of iterations to train for. It precedes num_epochs which means that
-# # if both are set, num_epochs will not be guaranteed.
-# # e.g., when 1 epoch is 1000 steps => `num_epochs: 2` and `max_steps: 100` will train for 100 steps
-# max_steps:
-
-# eval_table_size: # Approximate number of predictions sent to wandb depending on batch size. Enabled above 0. Default is 0
-# eval_table_max_new_tokens: # Total number of tokens generated for predictions sent to wandb. Default is 128
-
-# # Save model as safetensors (require safetensors package)
-# save_safetensors:
-
-# # Whether to mask out or include the human's prompt from the training labels
-# train_on_inputs: false
-# # Group similarly sized data to minimize padding.
-# # May be slower to start, as it must download and sort the entire dataset.
-# # Note that training loss may have an oscillating pattern with this enabled.
-# group_by_length: false
-
-# # Whether to use gradient checkpointing https://huggingface.co/docs/transformers/v4.18.0/en/performance#gradient-checkpointing
-# gradient_checkpointing: false
-
-# # Stop training after this many evaluation losses have increased in a row
-# # https://huggingface.co/transformers/v4.2.2/_modules/transformers/trainer_callback.html#EarlyStoppingCallback
-# early_stopping_patience: 3
-
-# # Specify a scheduler and kwargs to use with the optimizer
-# lr_scheduler: # 'one_cycle' | empty for cosine
-# lr_scheduler_kwargs:
-
-# # For one_cycle optim
-# lr_div_factor: # Learning rate div factor
-
-# # Specify optimizer
-# # Valid values are driven by the Transformers OptimizerNames class, see:
-# # https://github.com/huggingface/transformers/blob/95b374952dc27d8511541d6f5a4e22c9ec11fb24/src/transformers/training_args.py#L134
-# #
-# # Note that not all optimizers may be available in your environment, ex: 'adamw_anyprecision' is part of
-# # torchdistx, 'adamw_bnb_8bit' is part of bnb.optim.Adam8bit, etc. When in doubt, it is recommended to start with the optimizer used
-# # in the examples/ for your model and fine-tuning use case.
-# #
-# # Valid values for 'optimizer' include:
-# # - adamw_hf
-# # - adamw_torch
-# # - adamw_torch_fused
-# # - adamw_torch_xla
-# # - adamw_apex_fused
-# # - adafactor
-# # - adamw_anyprecision
-# # - sgd
-# # - adagrad
-# # - adamw_bnb_8bit
-# # - lion_8bit
-# # - lion_32bit
-# # - paged_adamw_32bit
-# # - paged_adamw_8bit
-# # - paged_lion_32bit
-# # - paged_lion_8bit
-# optimizer:
-# # Specify weight decay
-# weight_decay:
-# # adamw hyperparams
-# adam_beta1:
-# adam_beta2:
-# adam_epsilon:
-# # Gradient clipping max norm
-# max_grad_norm:
-
-# # Augmentation techniques
-# # NEFT https://arxiv.org/abs/2310.05914, set this to a number (paper default is 5) to add noise to embeddings
-# # currently only supported on Llama and Mistral
-# noisy_embedding_alpha:
-
-# # Whether to bettertransformers
-# flash_optimum:
-# # Whether to use xformers attention patch https://github.com/facebookresearch/xformers:
-# xformers_attention:
-# # Whether to use flash attention patch https://github.com/Dao-AILab/flash-attention:
-# flash_attention:
-# flash_attn_cross_entropy:  # Whether to use flash-attention cross entropy implementation - advanced use only
-# flash_attn_rms_norm:  # Whether to use flash-attention rms norm implementation - advanced use only
-# flash_attn_fuse_mlp: # Whether to fuse part of the MLP into a single operation
-# # Whether to use scaled-dot-product attention
-# # https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html
-# sdp_attention:
-# # Landmark attention (only llama)
-# landmark_attention:
-# # xpos RoPE see https://github.com/kaiokendev/cutoff-len-is-context-len/blob/main/util/xpos_rope_llama_monkey_patch.py
-# # LLaMA only
-# xpos_rope:
-
-# # Resume from a specific checkpoint dir
-# resume_from_checkpoint:
-# # If resume_from_checkpoint isn't set and you simply want it to start where it left off.
-# # Be careful with this being turned on between different models.
-# auto_resume_from_checkpoints: false
-
-# # Don't mess with this, it's here for accelerate and torchrun
-# local_rank:
-
-# # Add or change special tokens.
-# # If you add tokens here, you don't need to add them to the `tokens` list.
-# special_tokens:
-#   # bos_token: "<s>"
-#   # eos_token: "</s>"
-#   # unk_token: "<unk>"
-
-# # Add extra tokens.
-# tokens:
-
-# # FSDP
-# fsdp:
-# fsdp_config:
-
-# # Deepspeed config path. e.g., deepspeed/zero3.json
-# deepspeed:
-
-# # Advanced DDP Arguments
-# ddp_timeout:
-# ddp_bucket_cap_mb:
-# ddp_broadcast_buffers:
-
-# # Path to torch distx for optim 'adamw_anyprecision'
-# torchdistx_path:
-
-# # Set to HF dataset for type: 'completion' for streaming instead of pre-tokenize
-# pretraining_dataset:
-
-# # Debug mode
-# debug:
-
-# # Seed
-# seed:
-
-# # Allow overwrite yml config using from cli
-# strict:
-
-
-
-base_model: ${BASE_MODEL}
-base_model_ignore_patterns: ${BASE_MODEL_IGNORE_PATTERNS}
-base_model_config: ${BASE_MODEL_CONFIG}
-revision_of_model: ${REVISION_OF_MODEL}
-tokenizer_config: ${TOKENIZER_CONFIG}
-model_type: ${MODEL_TYPE}
-tokenizer_type: ${TOKENIZER_TYPE}
-trust_remote_code: ${TRUST_REMOTE_CODE}
-tokenizer_use_fast: ${TOKENIZER_USE_FAST}
-tokenizer_legacy: ${TOKENIZER_LEGACY}
-resize_token_embeddings_to_32x: ${RESIZE_TOKEN_EMBEDDINGS_TO_32X}
-
-is_falcon_derived_model: ${IS_FALCON_DERIVED_MODEL}
-is_llama_derived_model: ${IS_LLAMA_DERIVED_MODEL}
-is_qwen_derived_model: ${IS_QWEN_DERIVED_MODEL}
-is_mistral_derived_model: ${IS_MISTRAL_DERIVED_MODEL}
-
-overrides_of_model_config:
-  rope_scaling:
-    type: ${ROPE_SCALING_TYPE}
-    factor: ${ROPE_SCALING_FACTOR}
-
-bnb_config_kwargs:
-  llm_int8_has_fp16_weight: ${BNB_LLM_INT8_HAS_FP16_WEIGHT}
-  bnb_4bit_quant_type: ${BNB_4BIT_QUANT_TYPE}
-  bnb_4bit_use_double_quant: ${BNB_4BIT_USE_DOUBLE_QUANT}
-
-gptq: ${GPTQ}
-load_in_8bit: ${LOAD_IN_8BIT}
-load_in_4bit: ${LOAD_IN_4BIT}
-bf16: ${BF16}
-fp16: ${FP16}
-tf32: ${TF32}
-bfloat16: ${BFLOAT16}
-float16: ${FLOAT16}
-
-gpu_memory_limit: ${GPU_MEMORY_LIMIT}
-lora_on_cpu: ${LORA_ON_CPU}
-
-datasets:
-  - path: ${DATASET_PATH}
-    type: ${DATASET_TYPE}
-    ds_type: ${DATASET_DS_TYPE}
-    data_files: ${DATASET_DATA_FILES}
-    shards: ${DATASET_SHARDS}
-    name: ${DATASET_NAME}
-    train_on_split: ${DATASET_TRAIN_ON_SPLIT}
-    revision: ${DATASET_REVISION}
-    trust_remote_code: ${DATASET_TRUST_REMOTE_CODE}
-
-rl: ${RL}
-dpo_use_weighting: ${DPO_USE_WEIGHTING}
-
-chat_template: ${CHAT_TEMPLATE}
-chat_template_jinja: ${CHAT_TEMPLATE_JINJA}
-default_system_message: ${DEFAULT_SYSTEM_MESSAGE}
-dataset_prepared_path: ${DATASET_PREPARED_PATH}
-push_dataset_to_hub: ${PUSH_DATASET_TO_HUB}
-dataset_processes: ${DATASET_PROCESSES}
-dataset_keep_in_memory: ${DATASET_KEEP_IN_MEMORY}
-hub_model_id: ${HUB_MODEL_ID}
-hub_strategy: ${HUB_STRATEGY}
-hf_use_auth_token: ${HF_USE_AUTH_TOKEN}
-val_set_size: ${VAL_SET_SIZE}
-dataset_shard_num: ${DATASET_SHARD_NUM}
-dataset_shard_idx: ${DATASET_SHARD_IDX}
-
-sequence_len: ${SEQUENCE_LEN}
-pad_to_sequence_len: ${PAD_TO_SEQUENCE_LEN}
-sample_packing: ${SAMPLE_PACKING}
-eval_sample_packing: ${EVAL_SAMPLE_PACKING}
-sample_packing_eff_est: ${SAMPLE_PACKING_EFF_EST}
-total_num_tokens: ${TOTAL_NUM_TOKENS}
-sample_packing_group_size: ${SAMPLE_PACKING_GROUP_SIZE}
-sample_packing_bin_size: ${SAMPLE_PACKING_BIN_SIZE}
-
-batch_flattening: ${BATCH_FLATTENING}
-device_map: ${DEVICE_MAP}
-max_memory: ${MAX_MEMORY}
-
-adapter: ${ADAPTER}
-lora_model_dir: ${LORA_MODEL_DIR}
-
-lora_r: ${LORA_R}
-lora_alpha: ${LORA_ALPHA}
-lora_dropout: ${LORA_DROPOUT}
-lora_target_modules:
-  - ${LORA_TARGET_MODULES}
-lora_target_linear: ${LORA_TARGET_LINEAR}
-peft_layers_to_transform: ${PEFT_LAYERS_TO_TRANSFORM}
-lora_modules_to_save: ${LORA_MODULES_TO_SAVE}
-lora_fan_in_fan_out: ${LORA_FAN_IN_FAN_OUT}
-
-loraplus_lr_ratio: ${LORAPLUS_LR_RATIO}
-loraplus_lr_embedding: ${LORAPLUS_LR_EMBEDDING}
-
-peft:
-  loftq_config:
-    loftq_bits: ${LOFTQ_BITS}
-
-relora_steps: ${RELORA_STEPS}
-relora_warmup_steps: ${RELORA_WARMUP_STEPS}
-relora_anneal_steps: ${RELORA_ANNEAL_STEPS}
-relora_prune_ratio: ${RELORA_PRUNE_RATIO}
-relora_cpu_offload: ${RELORA_CPU_OFFLOAD}
-
-wandb_mode: ${WANDB_MODE}
-wandb_project: ${WANDB_PROJECT}
-wandb_entity: ${WANDB_ENTITY}
-wandb_watch: ${WANDB_WATCH}
-wandb_name: ${WANDB_NAME}
-wandb_run_id: ${WANDB_RUN_ID}
-wandb_log_model: ${WANDB_LOG_MODEL}
-
-mlflow_tracking_uri: ${MLFLOW_TRACKING_URI}
-mlflow_experiment_name: ${MLFLOW_EXPERIMENT_NAME}
-mlflow_run_name: ${MLFLOW_RUN_NAME}
-hf_mlflow_log_artifacts: ${HF_MLFLOW_LOG_ARTIFACTS}
-
-use_comet: ${USE_COMET}
-comet_api_key: ${COMET_API_KEY}
-comet_workspace: ${COMET_WORKSPACE}
-comet_project_name: ${COMET_PROJECT_NAME}
-comet_experiment_key: ${COMET_EXPERIMENT_KEY}
-comet_mode: ${COMET_MODE}
-comet_online: ${COMET_ONLINE}
-comet_experiment_config: ${COMET_EXPERIMENT_CONFIG}
-
-output_dir: ${OUTPUT_DIR}
-
-torch_compile: ${TORCH_COMPILE}
-torch_compile_backend: ${TORCH_COMPILE_BACKEND}
-
-gradient_accumulation_steps: ${GRADIENT_ACCUMULATION_STEPS}
-micro_batch_size: ${MICRO_BATCH_SIZE}
-eval_batch_size: ${EVAL_BATCH_SIZE}
-num_epochs: ${NUM_EPOCHS}
-warmup_steps: ${WARMUP_STEPS}
-warmup_ratio: ${WARMUP_RATIO}
-learning_rate: ${LEARNING_RATE}
-lr_quadratic_warmup: ${LR_QUADRATIC_WARMUP}
-logging_steps: ${LOGGING_STEPS}
-eval_steps: ${EVAL_STEPS}
-evals_per_epoch: ${EVALS_PER_EPOCH}
-save_strategy: ${SAVE_STRATEGY}
-save_steps: ${SAVE_STEPS}
-saves_per_epoch: ${SAVES_PER_EPOCH}
-save_total_limit: ${SAVE_TOTAL_LIMIT}
-max_steps: ${MAX_STEPS}
-
-eval_table_size: ${EVAL_TABLE_SIZE}
-eval_max_new_tokens: ${EVAL_MAX_NEW_TOKENS}
-eval_causal_lm_metrics: ${EVAL_CAUSAL_LM_METRICS}
-
-profiler_steps: ${PROFILER_STEPS}
-loss_watchdog_threshold: ${LOSS_WATCHDOG_THRESHOLD}
-loss_watchdog_patience: ${LOSS_WATCHDOG_PATIENCE}
-
-save_safetensors: ${SAVE_SAFETENSORS}
-train_on_inputs: ${TRAIN_ON_INPUTS}
-group_by_length: ${GROUP_BY_LENGTH}
-gradient_checkpointing: ${GRADIENT_CHECKPOINTING}
-early_stopping_patience: ${EARLY_STOPPING_PATIENCE}
-
-lr_scheduler: ${LR_SCHEDULER}
-lr_scheduler_kwargs: ${LR_SCHEDULER_KWARGS}
-cosine_min_lr_ratio: ${COSINE_MIN_LR_RATIO}
-cosine_constant_lr_ratio: ${COSINE_CONSTANT_LR_RATIO}
-lr_div_factor: ${LR_DIV_FACTOR}
-
-optimizer: ${OPTIMIZER}
-optim_args: ${OPTIM_ARGS}
-optim_target_modules: ${OPTIM_TARGET_MODULES}
-weight_decay: ${WEIGHT_DECAY}
-adam_beta1: ${ADAM_BETA1}
-adam_beta2: ${ADAM_BETA2}
-adam_epsilon: ${ADAM_EPSILON}
-max_grad_norm: ${MAX_GRAD_NORM}
-
-neftune_noise_alpha: ${NEFTUNE_NOISE_ALPHA}
-
-flash_optimum: ${FLASH_OPTIMUM}
-xformers_attention: ${XFORMERS_ATTENTION}
-flash_attention: ${FLASH_ATTENTION}
-flash_attn_cross_entropy: ${FLASH_ATTN_CROSS_ENTROPY}
-flash_attn_rms_norm: ${FLASH_ATTN_RMS_NORM}
-flash_attn_fuse_mlp: ${FLASH_ATTN_FUSE_MLP}
-sdp_attention: ${SDP_ATTENTION}
-s2_attention: ${S2_ATTENTION}
-resume_from_checkpoint: ${RESUME_FROM_CHECKPOINT}
-auto_resume_from_checkpoints: ${AUTO_RESUME_FROM_CHECKPOINTS}
-
-local_rank: ${LOCAL_RANK}
-
-special_tokens:
-  bos_token: ${SPECIAL_TOKEN_BOS}
-  eos_token: ${SPECIAL_TOKEN_EOS}
-  unk_token: ${SPECIAL_TOKEN_UNK}
-  pad_token: ${SPECIAL_TOKEN_PAD}
-
-tokens: ${TOKENS}
-
-fsdp: ${FSDP}
-fsdp_config: ${FSDP_CONFIG}
-deepspeed: ${DEEPSPEED}
-
-ddp_timeout: ${DDP_TIMEOUT}
-ddp_bucket_cap_mb: ${DDP_BUCKET_CAP_MB}
-ddp_broadcast_buffers: ${DDP_BROADCAST_BUFFERS}
-
-torchdistx_path: ${TORCHDISTX_PATH}
-pretraining_dataset: ${PRETRAINING_DATASET}
-debug: ${DEBUG}
-seed: ${SEED}
-strict: ${STRICT}
--- a/.runpod/src/handler.py
+++ b/.runpod/src/handler.py
@@ -1,66 +0,0 @@
-"""
-Runpod serverless entrypoint handler
-"""
-
-import os
-
-import runpod
-import yaml
-from huggingface_hub._login import login
-from train import train
-from utils import get_output_dir
-
-BASE_VOLUME = os.environ.get("BASE_VOLUME", "/runpod-volume")
-if not os.path.exists(BASE_VOLUME):
-    os.makedirs(BASE_VOLUME)
-
-logger = runpod.RunPodLogger()
-
-
-async def handler(job):
-    runpod_job_id = job["id"]
-    inputs = job["input"]
-    run_id = inputs.get("run_id", "default_run_id")
-    args = inputs.get("args", {})
-
-    # Set output directory
-    output_dir = os.path.join(BASE_VOLUME, get_output_dir(run_id))
-    args["output_dir"] = output_dir
-
-    # First save args to a temporary config file
-    config_path = "/workspace/test_config.yaml"
-
-    # Add run_name and job_id to args before saving
-    args["run_name"] = run_id
-    args["runpod_job_id"] = runpod_job_id
-
-    yaml_data = yaml.dump(args, default_flow_style=False)
-    with open(config_path, "w", encoding="utf-8") as file:
-        file.write(yaml_data)
-
-    # Handle credentials
-    credentials = inputs.get("credentials", {})
-
-    if "wandb_api_key" in credentials:
-        os.environ["WANDB_API_KEY"] = credentials["wandb_api_key"]
-    if "hf_token" in credentials:
-        os.environ["HF_TOKEN"] = credentials["hf_token"]
-
-    if os.environ.get("HF_TOKEN"):
-        login(token=os.environ["HF_TOKEN"])
-    else:
-        logger.info("No HF_TOKEN provided. Skipping login.")
-
-    logger.info("Starting Training.")
-    async for result in train(config_path):  # Pass the config path instead of args
-        logger.info(result)
-    logger.info("Training Complete.")
-
-    # Cleanup
-    if "WANDB_API_KEY" in os.environ:
-        del os.environ["WANDB_API_KEY"]
-    if "HF_TOKEN" in os.environ:
-        del os.environ["HF_TOKEN"]
-
-
-runpod.serverless.start({"handler": handler, "return_aggregate_stream": True})
--- a/.runpod/src/test_input.json
+++ b/.runpod/src/test_input.json
@@ -1,61 +0,0 @@
-{
-  "input": {
-    "user_id": "user",
-    "model_id": "llama-test",
-    "run_id": "llama-test",
-    "credentials": {
-      "wandb_api_key": "",
-      "hf_token": ""
-    },
-    "args": {
-      "base_model": "NousResearch/Meta-Llama-3-8B",
-      "model_type": "LlamaForCausalLM",
-      "tokenizer_type": "AutoTokenizer",
-      "load_in_8bit": true,
-      "load_in_4bit": false,
-      "strict": false,
-      "datasets": [
-        {
-          "path": "mhenrichsen/alpaca_2k_test",
-          "type": "alpaca"
-        }
-      ],
-      "val_set_size": 0.05,
-      "output_dir": "./outputs/lora-out",
-      "sequence_len": 4096,
-      "sample_packing": true,
-      "eval_sample_packing": false,
-      "pad_to_sequence_len": true,
-      "adapter": "lora",
-      "lora_r": 32,
-      "lora_alpha": 16,
-      "lora_dropout": 0.05,
-      "lora_target_linear": true,
-      "lora_modules_to_save": [
-        "embed_tokens",
-        "lm_head"
-      ],
-      "gradient_accumulation_steps": 4,
-      "micro_batch_size": 2,
-      "num_epochs": 1,
-      "optimizer": "adamw_bnb_8bit",
-      "lr_scheduler": "cosine",
-      "learning_rate": 0.0002,
-      "train_on_inputs": false,
-      "group_by_length": false,
-      "bf16": "auto",
-      "tf32": false,
-      "gradient_checkpointing": true,
-      "logging_steps": 1,
-      "flash_attention": true,
-      "warmup_steps": 1,
-      "evals_per_epoch": 1,
-      "eval_max_new_tokens": 128,
-      "saves_per_epoch": 1,
-      "weight_decay": 0.0,
-      "special_tokens": {
-        "pad_token": "<|end_of_text|>"
-      }
-    }
-  }
-}
--- a/.runpod/src/train.py
+++ b/.runpod/src/train.py
@@ -1,45 +0,0 @@
-"""
-Runpod train entrypoint
-"""
-
-import asyncio
-
-
-async def train(config_path: str, gpu_id: str = "0", preprocess: bool = True):
-    """
-    Run preprocessing (if enabled) and training with the given config file
-    :param config_path: Path to the YAML config file
-    :param gpu_id: GPU ID to use (default: "0")
-    :param preprocess: Whether to run preprocessing (default: True)
-
-    """
-    # First check if preprocessing is needed
-    if preprocess:
-        # Preprocess command
-        preprocess_cmd = (
-            f"CUDA_VISIBLE_DEVICES={gpu_id} axolotl preprocess {config_path}"
-        )
-        process = await asyncio.create_subprocess_shell(
-            preprocess_cmd,
-            stdout=asyncio.subprocess.PIPE,
-            stderr=asyncio.subprocess.STDOUT,
-        )
-
-        if process.stdout is not None:
-            async for line in process.stdout:
-                yield f"Preprocessing: {line.decode().strip()}"
-        await process.wait()
-        yield "Preprocessing completed."
-    else:
-        yield "Skipping preprocessing step."
-
-    # Training command
-    train_cmd = f"axolotl train {config_path}"
-    process = await asyncio.create_subprocess_shell(
-        train_cmd, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.STDOUT
-    )
-
-    if process.stdout is not None:
-        async for line in process.stdout:
-            yield f"Training: {line.decode().strip()}"
-    await process.wait()
--- a/.runpod/src/utils.py
+++ b/.runpod/src/utils.py
@@ -1,89 +0,0 @@
-"""
-Runpod launcher utils
-"""
-
-import os
-
-import yaml
-
-
-def get_output_dir(run_id):
-    path = f"fine-tuning/{run_id}"
-    return path
-
-
-def make_valid_config(input_args):
-    """
-    Creates and saves updated config file, returns the path to the new config
-    :param input_args: dict of input args
-    :return: str, path to the updated config file
-    """
-    # Load default config
-    with open("config/config.yaml", "r", encoding="utf-8") as fin:
-        all_args = yaml.safe_load(fin)
-
-    if not input_args:
-        print("No args provided, using defaults")
-    else:
-        all_args.update(input_args)
-
-    # Create updated config path
-    updated_config_path = "config/updated_config.yaml"
-
-    # Save updated config to new file
-    with open(updated_config_path, "w", encoding="utf-8") as f:
-        yaml.dump(all_args, f)
-
-    return updated_config_path
-
-
-def set_config_env_vars(args: dict):
-    """
-    Convert API arguments into environment variables.
-    Handles nested dictionaries, lists, and special values.
-
-    Args:
-        args (dict): The arguments dictionary from the API request
-    """
-
-    def process_value(value):
-        """Convert Python values to string format for environment variables"""
-        if value is None:
-            return ""
-        if isinstance(value, bool):
-            return str(value).lower()
-        if isinstance(value, (list, dict)):
-            return str(value)
-        return str(value)
-
-    def set_env_vars(data, prefix=""):
-        """Recursively set environment variables from nested dictionary"""
-        for key, value in data.items():
-            env_key = prefix + key.upper()
-
-            # Handle special cases
-            if isinstance(value, dict):
-                # For nested dictionaries (like special_tokens)
-                set_env_vars(value, f"{env_key}_")
-            elif isinstance(value, list):
-                # Handle list of dictionaries (like datasets)
-                if value and isinstance(value[0], dict):
-                    for i, item in enumerate(value):
-                        set_env_vars(item, f"{env_key}_{i}_")
-                else:
-                    # For simple lists (like lora_target_modules)
-                    os.environ[env_key] = process_value(value)
-            else:
-                # Handle all other cases
-                os.environ[env_key] = process_value(value)
-
-    # Clear any existing related environment variables
-    # This prevents old values from persisting
-    for key in list(os.environ.keys()):
-        if key.startswith(
-            ("BASE_MODEL", "MODEL_TYPE", "TOKENIZER_TYPE", "DATASET", "LORA_", "WANDB_")
-        ):
-            del os.environ[key]
-
-    # Set new environment variables
-    set_env_vars(args)
--- a/.runpod/test-input.json
+++ b/.runpod/test-input.json
@@ -1,86 +0,0 @@
-{
-  "input": {
-    "name": "quick_smoke_test_sft",
-    "user_id": "user",
-    "model_id": "llama-test",
-    "run_id": "llama-test",
-    "credentials": {
-      "wandb_api_key": "",
-      "hf_token": ""
-    },
-    "args": {
-      "base_model": "HuggingFaceTB/SmolLM2-135M",
-      "model_type": "AutoModelForCausalLM",
-      "tokenizer_type": "AutoTokenizer",
-      "load_in_4bit": true,
-      "strict": false,
-      "datasets": [
-        {
-          "path": "mhenrichsen/alpaca_2k_test",
-          "type": "alpaca",
-          "split": "train[:10%]"
-        }
-      ],
-      "val_set_size": 0.02,
-      "output_dir": "./outputs/lora-out",
-      "sequence_len": 4096,
-      "sample_packing": true,
-      "eval_sample_packing": false,
-      "pad_to_sequence_len": true,
-      "adapter": "qlora",
-      "lora_r": 32,
-      "lora_alpha": 64,
-      "lora_dropout": 0.05,
-      "lora_target_linear": true,
-      "lora_modules_to_save": [
-        "embed_tokens",
-        "lm_head"
-      ],
-      "gradient_accumulation_steps": 2,
-      "micro_batch_size": 1,
-      "num_epochs": 1,
-      "optimizer": "adamw_torch_fused",
-      "lr_scheduler": "cosine",
-      "learning_rate": 0.0002,
-      "train_on_inputs": false,
-      "group_by_length": false,
-      "bf16": "auto",
-      "tf32": true,
-      "gradient_checkpointing": true,
-      "logging_steps": 1,
-      "flash_attention": true,
-      "warmup_steps": 1,
-      "evals_per_epoch": 1,
-      "eval_max_new_tokens": 128,
-      "saves_per_epoch": 1,
-      "weight_decay": 0.0,
-      "special_tokens": {
-        "pad_token": "<|endoftext|>"
-      },
-      "max_steps": 20
-    },
-    "timeout": 100000
-  },
-  "config": {
-    "gpuTypeId": "NVIDIA GeForce RTX 4090",
-    "gpuCount": 1,
-    "containerDiskInGb": 200,
-    "env": [
-      {
-        "key": "TOKENIZER",
-        "value": ""
-      },
-      {
-        "key": "DISABLE_LOG_STATS",
-        "value": "true"
-      }
-    ],
-    "allowedCudaVersions": [
-      "12.8",
-      "12.7",
-      "12.6",
-      "12.5",
-      "12.4"
-    ]
-  }
-}
--- a/.runpod/tests.json
+++ b/.runpod/tests.json
@@ -1,90 +0,0 @@
-{
-  "tests": [
-    {
-      "name": "quick_smoke_test_sft",
-      "input": {
-        "user_id": "user",
-        "model_id": "llama-test",
-        "run_id": "llama-test",
-        "credentials": {
-          "wandb_api_key": "",
-          "hf_token": ""
-        },
-        "args": {
-          "base_model": "HuggingFaceTB/SmolLM2-135M",
-          "model_type": "AutoModelForCausalLM",
-          "tokenizer_type": "AutoTokenizer",
-          "load_in_4bit": true,
-          "strict": false,
-          "datasets": [
-            {
-              "path": "mhenrichsen/alpaca_2k_test",
-              "type": "alpaca",
-              "split": "train[:10%]"
-            }
-          ],
-          "val_set_size": 0.02,
-          "output_dir": "./outputs/lora-out",
-          "sequence_len": 4096,
-          "sample_packing": true,
-          "eval_sample_packing": false,
-          "pad_to_sequence_len": true,
-          "adapter": "qlora",
-          "lora_r": 32,
-          "lora_alpha": 64,
-          "lora_dropout": 0.05,
-          "lora_target_linear": true,
-          "lora_modules_to_save": [
-            "embed_tokens",
-            "lm_head"
-          ],
-          "gradient_accumulation_steps": 2,
-          "micro_batch_size": 1,
-          "num_epochs": 1,
-          "optimizer": "adamw_torch_fused",
-          "lr_scheduler": "cosine",
-          "learning_rate": 0.0002,
-          "train_on_inputs": false,
-          "group_by_length": false,
-          "bf16": "auto",
-          "tf32": true,
-          "gradient_checkpointing": true,
-          "logging_steps": 1,
-          "flash_attention": true,
-          "warmup_steps": 1,
-          "evals_per_epoch": 1,
-          "eval_max_new_tokens": 128,
-          "saves_per_epoch": 1,
-          "weight_decay": 0.0,
-          "special_tokens": {
-            "pad_token": "<|endoftext|>"
-          },
-          "max_steps": 20
-        }
-      },
-      "timeout": 100000
-    }
-  ],
-  "config": {
-    "gpuTypeId": "NVIDIA GeForce RTX 4090",
-    "gpuCount": 1,
-    "containerDiskInGb": 200,
-    "env": [
-      {
-        "key": "TOKENIZER",
-        "value": ""
-      },
-      {
-        "key": "DISABLE_LOG_STATS",
-        "value": "true"
-      }
-    ],
-    "allowedCudaVersions": [
-      "12.8",
-      "12.7",
-      "12.6",
-      "12.5",
-      "12.4"
-    ]
-  }
-}
--- a/CITATION.cff
+++ b/CITATION.cff
@@ -1,10 +0,0 @@
-cff-version: 1.2.0
-type: software
-title: "Axolotl: Post-Training for AI Models"
-message: "If you use this software, please cite it as below."
-authors:
-  - name: "Axolotl maintainers and contributors"
-repository-code: "https://github.com/axolotl-ai-cloud/axolotl"
-url: "https://axolotl.ai/"
-license: Apache-2.0
-date-released: "2023-05-30"
--- a/1
+++ b/1
@@ -1 +0,0 @@
-docs.axolotl.ai
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -2,5 +2,4 @@ include requirements.txt
 include README.md
 include LICENSE
 include src/setuptools_axolotl_dynamic_dependencies.py
-include src/axolotl/utils/chat_templates/templates/*.jinja
 recursive-include axolotl *.py
--- a/README.md
+++ b/README.md
@@ -1,15 +1,14 @@
 <p align="center">
    <picture>
-        <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/887513285d98132142bf5db2a74eb5e0928787f1/image/axolotl_logo_digital_white.svg">
-        <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/887513285d98132142bf5db2a74eb5e0928787f1/image/axolotl_logo_digital_black.svg">
-        <img alt="Axolotl" src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/887513285d98132142bf5db2a74eb5e0928787f1/image/axolotl_logo_digital_black.svg" width="400" height="104" style="max-width: 100%;">
+        <source media="(prefers-color-scheme: dark)" srcset="image/axolotl_logo_digital_white.svg">
+        <source media="(prefers-color-scheme: light)" srcset="image/axolotl_logo_digital_black.svg">
+        <img alt="Axolotl" src="image/axolotl_logo_digital_black.svg" width="400" height="104" style="max-width: 100%;">
    </picture>
 </p>

 <p align="center">
    <img src="https://img.shields.io/github/license/axolotl-ai-cloud/axolotl.svg?color=blue" alt="GitHub License">
    <img src="https://github.com/axolotl-ai-cloud/axolotl/actions/workflows/tests.yml/badge.svg" alt="tests">
-    <a href="https://codecov.io/gh/axolotl-ai-cloud/axolotl"><img src="https://codecov.io/gh/axolotl-ai-cloud/axolotl/branch/main/graph/badge.svg" alt="codecov"></a>
    <a href="https://github.com/axolotl-ai-cloud/axolotl/releases"><img src="https://img.shields.io/github/release/axolotl-ai-cloud/axolotl.svg" alt="Releases"></a>
    <br/>
    <a href="https://github.com/axolotl-ai-cloud/axolotl/graphs/contributors"><img src="https://img.shields.io/github/contributors-anon/axolotl-ai-cloud/axolotl?color=yellow&style=flat-square" alt="contributors" style="height: 20px;"></a>
@@ -22,147 +21,774 @@
    <img src="https://github.com/axolotl-ai-cloud/axolotl/actions/workflows/multi-gpu-e2e.yml/badge.svg" alt="multigpu-semi-weekly tests">
 </p>

-
-## 🎉 Latest Updates
-
- 2025/07:
-  - ND Parallelism support has been added into Axolotl. Compose Context Parallelism (CP), Tensor Parallelism (TP), and Fully Sharded Data Parallelism (FSDP) within a single node and across multiple nodes. Check out the [blog post](https://huggingface.co/blog/accelerate-nd-parallel) for more info.
-  - Axolotl adds more models: [GPT-OSS](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/gpt-oss), [Gemma 3n](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/gemma3n), [Liquid Foundation Model 2 (LFM2)](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/lfm2), and [Arcee Foundation Models (AFM)](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/afm).
-  - FP8 finetuning with fp8 gather op is now possible in Axolotl via `torchao`. Get started [here](https://docs.axolotl.ai/docs/mixed_precision.html#sec-fp8)!
-  - [Voxtral](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/voxtral), [Magistral 1.1](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/magistral), and [Devstral](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/devstral) with mistral-common tokenizer support has been integrated in Axolotl!
-  - TiledMLP support for single-GPU to multi-GPU training with DDP, DeepSpeed and FSDP support has been added to support Arctic Long Sequence Training. (ALST). See [examples](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/alst) for using ALST with Axolotl!
- 2025/05: Quantization Aware Training (QAT) support has been added to Axolotl. Explore the [docs](https://docs.axolotl.ai/docs/qat.html) to learn more!
- 2025/03: Axolotl has implemented Sequence Parallelism (SP) support. Read the [blog](https://huggingface.co/blog/axolotl-ai-co/long-context-with-sequence-parallelism-in-axolotl) and [docs](https://docs.axolotl.ai/docs/sequence_parallelism.html) to learn how to scale your context length when fine-tuning.
-
-<details>
-
-<summary>Expand older updates</summary>
-
- 2025/06: Magistral with mistral-common tokenizer support has been added to Axolotl. See [examples](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/magistral) to start training your own Magistral models with Axolotl!
- 2025/04: Llama 4 support has been added in Axolotl. See [examples](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/llama-4) to start training your own Llama 4 models with Axolotl's linearized version!
- 2025/03: (Beta) Fine-tuning Multimodal models is now supported in Axolotl. Check out the [docs](https://docs.axolotl.ai/docs/multimodal.html) to fine-tune your own!
- 2025/02: Axolotl has added LoRA optimizations to reduce memory usage and improve training speed for LoRA and QLoRA in single GPU and multi-GPU training (DDP and DeepSpeed). Jump into the [docs](https://docs.axolotl.ai/docs/lora_optims.html) to give it a try.
- 2025/02: Axolotl has added GRPO support. Dive into our [blog](https://huggingface.co/blog/axolotl-ai-co/training-llms-w-interpreter-feedback-wasm) and [GRPO example](https://github.com/axolotl-ai-cloud/grpo_code) and have some fun!
- 2025/01: Axolotl has added Reward Modelling / Process Reward Modelling fine-tuning support. See [docs](https://docs.axolotl.ai/docs/reward_modelling.html).
-
-</details>
-
-## ✨ Overview
-
-Axolotl is a tool designed to streamline post-training for various AI models.
+Axolotl is a tool designed to streamline the fine-tuning of various AI models, offering support for multiple configurations and architectures.

 Features:
+- Train various Huggingface models such as llama, pythia, falcon, mpt
+- Supports fullfinetune, lora, qlora, relora, and gptq
+- Customize configurations using a simple yaml file or CLI overwrite
+- Load different dataset formats, use custom formats, or bring your own tokenized datasets
+- Integrated with xformer, flash attention, [liger kernel](https://github.com/linkedin/Liger-Kernel), rope scaling, and multipacking
+- Works with single GPU or multiple GPUs via FSDP or Deepspeed
+- Easily run with Docker locally or on the cloud
+- Log results and optionally checkpoints to wandb, mlflow or Comet
+- And more!

- **Multiple Model Support**: Train various models like LLaMA, Mistral, Mixtral, Pythia, and more. We are compatible with HuggingFace transformers causal language models.
- **Training Methods**: Full fine-tuning, LoRA, QLoRA, GPTQ, QAT, Preference Tuning (DPO, IPO, KTO, ORPO), RL (GRPO), Multimodal, and Reward Modelling (RM) / Process Reward Modelling (PRM).
- **Easy Configuration**: Re-use a single YAML file between dataset preprocess, training, evaluation, quantization, and inference.
- **Performance Optimizations**: [Multipacking](https://docs.axolotl.ai/docs/multipack.html), [Flash Attention](https://github.com/Dao-AILab/flash-attention), [Xformers](https://github.com/facebookresearch/xformers), [Flex Attention](https://pytorch.org/blog/flexattention/), [Liger Kernel](https://github.com/linkedin/Liger-Kernel), [Cut Cross Entropy](https://github.com/apple/ml-cross-entropy/tree/main), [Sequence Parallelism (SP)](https://docs.axolotl.ai/docs/sequence_parallelism.html), [LoRA optimizations](https://docs.axolotl.ai/docs/lora_optims.html), [Multi-GPU training (FSDP1, FSDP2, DeepSpeed)](https://docs.axolotl.ai/docs/multi-gpu.html), [Multi-node training (Torchrun, Ray)](https://docs.axolotl.ai/docs/multi-node.html), and many more!
- **Flexible Dataset Handling**: Load from local, HuggingFace, and cloud (S3, Azure, GCP, OCI) datasets.
- **Cloud Ready**: We ship [Docker images](https://hub.docker.com/u/axolotlai) and also [PyPI packages](https://pypi.org/project/axolotl/) for use on cloud platforms and local hardware.
+<a href="https://www.phorm.ai/query?projectId=e315ba4a-4e14-421f-ab05-38a1f9076f25">
+  <img alt="phorm.ai" src="https://img.shields.io/badge/Phorm-Ask_AI-%23F2777A.svg?&logo=data:image/svg+xml;base64,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">
+</a>

+<table>
+<tr>
+<td>

+## Table of Contents
+- [Axolotl](#axolotl)
+  - [Table of Contents](#table-of-contents)
+  - [Quickstart ⚡](#quickstart-)
+    - [Edge Builds](#edge-builds-)
+    - [Axolotl CLI Usage](#axolotl-cli-usage)
+  - [Badge ❤🏷️](#badge-️)
+  - [Contributing 🤝](#contributing-)
+  - [Sponsors 🤝❤](#sponsors-)
+  - [Axolotl supports](#axolotl-supports)
+  - [Advanced Setup](#advanced-setup)
+    - [Environment](#environment)
+      - [Docker](#docker)
+      - [Conda/Pip venv](#condapip-venv)
+      - [Cloud GPU](#cloud-gpu)
+      - [Bare Metal Cloud GPU](#bare-metal-cloud-gpu)
+        - [LambdaLabs](#lambdalabs)
+        - [GCP](#gcp)
+      - [Windows](#windows)
+      - [Mac](#mac)
+      - [Google Colab](#google-colab)
+      - [Launching on public clouds via SkyPilot](#launching-on-public-clouds-via-skypilot)
+      - [Launching on public clouds via dstack](#launching-on-public-clouds-via-dstack)
+    - [Dataset](#dataset)
+    - [Config](#config)
+      - [All Config Options](#all-config-options)
+    - [Train](#train)
+      - [Preprocess dataset](#preprocess-dataset)
+      - [Multi-GPU](#multi-gpu)
+        - [DeepSpeed](#deepspeed)
+        - [FSDP](#fsdp)
+        - [FSDP + QLoRA](#fsdp--qlora)
+        - [Weights \& Biases Logging](#weights--biases-logging)
+        - [Special Tokens](#special-tokens)
+      - [Liger Kernel](#liger-kernel)
+    - [Inference Playground](#inference-playground)
+    - [Merge LORA to base](#merge-lora-to-base)
+  - [Common Errors 🧰](#common-errors-)
+    - [Tokenization Mismatch b/w Inference \& Training](#tokenization-mismatch-bw-inference--training)
+  - [Debugging Axolotl](#debugging-axolotl)
+  - [Need help? 🙋](#need-help-)

-## 🚀 Quick Start
+</td>
+<td>

-**Requirements**:
+<div align="center">
+  <img src="image/axolotl_symbol_digital_white.svg" alt="axolotl" width="160">
+  <div>
+    <p>
+      <b>Axolotl provides a unified repository for fine-tuning <br />a variety of AI models with ease</b>
+    </p>
+    <p>
+      Go ahead and Axolotl questions!!
+    </p>
+    <img src="https://github.com/axolotl-ai-cloud/axolotl/actions/workflows/pre-commit.yml/badge.svg?branch=main" alt="pre-commit">
+    <img alt="PyTest Status" src="https://github.com/axolotl-ai-cloud/axolotl/actions/workflows/tests.yml/badge.svg?branch=main">
+  </div>
+</div>

- NVIDIA GPU (Ampere or newer for `bf16` and Flash Attention) or AMD GPU
- Python 3.11
- PyTorch ≥2.6.0
+</td>
+</tr>
+</table>

-### Installation
+## Quickstart ⚡

-#### Using pip
+Get started with Axolotl in just a few steps! This quickstart guide will walk you through setting up and running a basic fine-tuning task.
+
+**Requirements**: *Nvidia* GPU (Ampere architecture or newer for `bf16` and Flash Attention) or *AMD* GPU, Python >=3.10 and PyTorch >=2.3.1.

 ```bash
-pip3 install -U packaging==23.2 setuptools==75.8.0 wheel ninja
 pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]

-# Download example axolotl configs, deepspeed configs
+# download examples and optionally deepspeed configs to the local path
 axolotl fetch examples
 axolotl fetch deepspeed_configs  # OPTIONAL
-```

-#### Using Docker
-
-Installing with Docker can be less error prone than installing in your own environment.
-```bash
-docker run --gpus '"all"' --rm -it axolotlai/axolotl:main-latest
-```
-
-Other installation approaches are described [here](https://docs.axolotl.ai/docs/installation.html).
-
-#### Cloud Providers
-
-<details>
-
- [RunPod](https://runpod.io/gsc?template=v2ickqhz9s&ref=6i7fkpdz)
- [Vast.ai](https://cloud.vast.ai?ref_id=62897&template_id=bdd4a49fa8bce926defc99471864cace&utm_source=github&utm_medium=developer_community&utm_campaign=template_launch_axolotl&utm_content=readme)
- [PRIME Intellect](https://app.primeintellect.ai/dashboard/create-cluster?image=axolotl&location=Cheapest&security=Cheapest&show_spot=true)
- [Modal](https://www.modal.com?utm_source=github&utm_medium=github&utm_campaign=axolotl)
- [Novita](https://novita.ai/gpus-console?templateId=311)
- [JarvisLabs.ai](https://jarvislabs.ai/templates/axolotl)
- [Latitude.sh](https://latitude.sh/blueprint/989e0e79-3bf6-41ea-a46b-1f246e309d5c)
-
-</details>
-
-### Your First Fine-tune
-
-```bash
-# Fetch axolotl examples
-axolotl fetch examples
-
-# Or, specify a custom path
-axolotl fetch examples --dest path/to/folder
-
-# Train a model using LoRA
+# finetune using lora
 axolotl train examples/llama-3/lora-1b.yml
 ```

-That's it! Check out our [Getting Started Guide](https://docs.axolotl.ai/docs/getting-started.html) for a more detailed walkthrough.
+### Edge Builds 🏎️

+If you're looking for the latest features and updates between releases, you'll need to install
+from source.

-## 📚 Documentation
-
- [Installation Options](https://docs.axolotl.ai/docs/installation.html) - Detailed setup instructions for different environments
- [Configuration Guide](https://docs.axolotl.ai/docs/config-reference.html) - Full configuration options and examples
- [Dataset Loading](https://docs.axolotl.ai/docs/dataset_loading.html) - Loading datasets from various sources
- [Dataset Guide](https://docs.axolotl.ai/docs/dataset-formats/) - Supported formats and how to use them
- [Multi-GPU Training](https://docs.axolotl.ai/docs/multi-gpu.html)
- [Multi-Node Training](https://docs.axolotl.ai/docs/multi-node.html)
- [Multipacking](https://docs.axolotl.ai/docs/multipack.html)
- [API Reference](https://docs.axolotl.ai/docs/api/) - Auto-generated code documentation
- [FAQ](https://docs.axolotl.ai/docs/faq.html) - Frequently asked questions
-
-## 🤝 Getting Help
-
- Join our [Discord community](https://discord.gg/HhrNrHJPRb) for support
- Check out our [Examples](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/) directory
- Read our [Debugging Guide](https://docs.axolotl.ai/docs/debugging.html)
- Need dedicated support? Please contact [✉️wing@axolotl.ai](mailto:wing@axolotl.ai) for options
-
-## 🌟 Contributing
-
-Contributions are welcome! Please see our [Contributing Guide](https://github.com/axolotl-ai-cloud/axolotl/blob/main/.github/CONTRIBUTING.md) for details.
-
-## ❤️ Sponsors
-
-Interested in sponsoring? Contact us at [wing@axolotl.ai](mailto:wing@axolotl.ai)
-
-## 📝 Citing Axolotl
-
-If you use Axolotl in your research or projects, please cite it as follows:
-
-```bibtex
-@software{axolotl,
-  title = {Axolotl: Post-Training for AI Models},
-  author = {{Axolotl maintainers and contributors}},
-  url = {https://github.com/axolotl-ai-cloud/axolotl},
-  license = {Apache-2.0},
-  year = {2023}
-}
+```bash
+git clone https://github.com/axolotl-ai-cloud/axolotl.git
+cd axolotl
+pip3 install packaging ninja
+pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'
 ```

-## 📜 License
+### Axolotl CLI Usage
+We now support a new, more streamlined CLI using [click](https://click.palletsprojects.com/en/stable/).

-This project is licensed under the Apache 2.0 License - see the [LICENSE](LICENSE) file for details.
+```bash
+# preprocess datasets - optional but recommended
+CUDA_VISIBLE_DEVICES="0" axolotl preprocess examples/llama-3/lora-1b.yml
+
+# finetune lora
+axolotl train examples/llama-3/lora-1b.yml
+
+# inference
+axolotl inference examples/llama-3/lora-1b.yml \
+    --lora-model-dir="./outputs/lora-out"
+
+# gradio
+axolotl inference examples/llama-3/lora-1b.yml \
+    --lora-model-dir="./outputs/lora-out" --gradio
+
+# remote yaml files - the yaml config can be hosted on a public URL
+# Note: the yaml config must directly link to the **raw** yaml
+axolotl train https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/examples/llama-3/lora-1b.yml
+```
+
+We've also added a new command for fetching `examples` and `deepspeed_configs` to your
+local machine. This will come in handy when installing `axolotl` from PyPI.
+
+```bash
+# Fetch example YAML files (stores in "examples/" folder)
+axolotl fetch examples
+
+# Fetch deepspeed config files (stores in "deepspeed_configs/" folder)
+axolotl fetch deepspeed_configs
+
+# Optionally, specify a destination folder
+axolotl fetch examples --dest path/to/folder
+```
+
+### Legacy Usage
+<details>
+
+<summary>Click to Expand</summary>
+
+While the Axolotl CLI is the preferred method for interacting with axolotl, we
+still support the legacy `-m axolotl.cli.*` usage.
+
+```bash
+# preprocess datasets - optional but recommended
+CUDA_VISIBLE_DEVICES="0" python -m axolotl.cli.preprocess examples/llama-3/lora-1b.yml
+
+# finetune lora
+accelerate launch -m axolotl.cli.train examples/llama-3/lora-1b.yml
+
+# inference
+accelerate launch -m axolotl.cli.inference examples/llama-3/lora-1b.yml \
+    --lora_model_dir="./outputs/lora-out"
+
+# gradio
+accelerate launch -m axolotl.cli.inference examples/llama-3/lora-1b.yml \
+    --lora_model_dir="./outputs/lora-out" --gradio
+
+# remote yaml files - the yaml config can be hosted on a public URL
+# Note: the yaml config must directly link to the **raw** yaml
+accelerate launch -m axolotl.cli.train https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/examples/llama-3/lora-1b.yml
+```
+
+</details>
+
+## Badge ❤🏷️
+
+Building something cool with Axolotl? Consider adding a badge to your model card.
+
+```markdown
+[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
+```
+
+[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
+
+## Sponsors 🤝❤
+
+If you love axolotl, consider sponsoring the project by reaching out directly to [wing@axolotl.ai](mailto:wing@axolotl.ai).
+
+---
+
+- [Modal](https://modal.com/) Modal lets you run data/AI jobs in the cloud, by just writing a few lines of Python. Customers use Modal to deploy Gen AI models at large scale, fine-tune LLM models, run protein folding simulations, and much more.
+
+---
+
+## Contributing 🤝
+
+Please read the [contributing guide](./.github/CONTRIBUTING.md)
+
+Bugs? Please check the [open issues](https://github.com/axolotl-ai-cloud/axolotl/issues/bug) else create a new Issue.
+
+PRs are **greatly welcome**!
+
+Please run the quickstart instructions followed by the below to setup env:
+```bash
+pip3 install -r requirements-dev.txt -r requirements-tests.txt
+pre-commit install
+
+# test
+pytest tests/
+
+# optional: run against all files
+pre-commit run --all-files
+```
+
+Thanks to all of our contributors to date. Help drive open source AI progress forward by contributing to Axolotl.
+
+<a href="https://github.com/axolotl-ai-cloud/axolotl/graphs/contributors">
+  <img src="https://contrib.rocks/image?repo=openaccess-ai-collective/axolotl" alt="contributor chart by https://contrib.rocks"/>
+</a>
+
+## Axolotl supports
+
+|             | fp16/fp32 | lora | qlora | gptq | gptq w/flash attn | flash attn | xformers attn |
+|-------------|:----------|:-----|-------|------|-------------------|------------|--------------|
+| llama       | ✅         | ✅    | ✅     | ✅             | ✅                 | ✅          | ✅            |
+| Mistral     | ✅         | ✅    | ✅     | ✅             | ✅                 | ✅          | ✅            |
+| Mixtral-MoE | ✅         | ✅    | ✅     | ❓             | ❓                 | ❓          | ❓            |
+| Mixtral8X22 | ✅         | ✅    | ✅     | ❓             | ❓                 | ❓          | ❓            |
+| Pythia      | ✅         | ✅    | ✅     | ❌             | ❌                 | ❌          | ❓            |
+| cerebras    | ✅         | ✅    | ✅     | ❌             | ❌                 | ❌          | ❓            |
+| btlm        | ✅         | ✅    | ✅     | ❌             | ❌                 | ❌          | ❓            |
+| mpt         | ✅         | ❌    | ❓     | ❌             | ❌                 | ❌          | ❓            |
+| falcon      | ✅         | ✅    | ✅     | ❌             | ❌                 | ❌          | ❓            |
+| gpt-j       | ✅         | ✅    | ✅     | ❌             | ❌                 | ❓          | ❓            |
+| XGen        | ✅         | ❓    | ✅     | ❓             | ❓                 | ❓          | ✅            |
+| phi         | ✅         | ✅    | ✅     | ❓             | ❓                 | ❓          | ❓            |
+| RWKV        | ✅         | ❓    | ❓     | ❓             | ❓                 | ❓          | ❓            |
+| Qwen        | ✅         | ✅    | ✅     | ❓             | ❓                 | ❓          | ❓            |
+| Gemma       | ✅         | ✅    | ✅     | ❓             | ❓                 | ✅          | ❓            |
+| Jamba       | ✅         | ✅    | ✅     | ❓             | ❓                 | ✅          | ❓            |
+
+✅: supported
+❌: not supported
+❓: untested
+
+## Advanced Setup
+
+### Environment
+
+#### Docker
+
+  ```bash
+  docker run --gpus '"all"' --rm -it axolotlai/axolotl:main-latest
+  ```
+
+  Or run on the current files for development:
+
+  ```sh
+  docker compose up -d
+  ```
+
+>[!Tip]
+> If you want to debug axolotl or prefer to use Docker as your development environment, see the [debugging guide's section on Docker](docs/debugging.qmd#debugging-with-docker).
+
+  <details>
+
+  <summary>Docker advanced</summary>
+
+  A more powerful Docker command to run would be this:
+
+  ```bash
+docker run --privileged --gpus '"all"' --shm-size 10g --rm -it --name axolotl --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 --mount type=bind,src="${PWD}",target=/workspace/axolotl -v ${HOME}/.cache/huggingface:/root/.cache/huggingface axolotlai/axolotl:main-latest
+  ```
+
+  It additionally:
+  * Prevents memory issues when running e.g. deepspeed (e.g. you could hit SIGBUS/signal 7 error) through `--ipc` and `--ulimit` args.
+  * Persists the downloaded HF data (models etc.) and your modifications to axolotl code through `--mount`/`-v` args.
+  * The `--name` argument simply makes it easier to refer to the container in vscode (`Dev Containers: Attach to Running Container...`) or in your terminal.
+  * The `--privileged` flag gives all capabilities to the container.
+  * The `--shm-size 10g` argument increases the shared memory size. Use this if you see `exitcode: -7` errors using deepspeed.
+
+  [More information on nvidia website](https://docs.nvidia.com/deeplearning/frameworks/user-guide/index.html#setincshmem)
+
+  </details>
+
+#### Conda/Pip venv
+  1. Install python >=**3.10**
+
+  2. Install pytorch stable https://pytorch.org/get-started/locally/
+
+  3. Install Axolotl along with python dependencies
+        ```bash
+        pip3 install packaging
+        pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'
+        ```
+  4. (Optional) Login to Huggingface to use gated models/datasets.
+        ```bash
+        huggingface-cli login
+        ```
+        Get the token at huggingface.co/settings/tokens
+
+#### Cloud GPU
+
+For cloud GPU providers that support docker images, use [`axolotlai/axolotl-cloud:main-latest`](https://hub.docker.com/r/axolotlai/axolotl-cloud/tags)
+
+- on Latitude.sh use this [direct link](https://latitude.sh/blueprint/989e0e79-3bf6-41ea-a46b-1f246e309d5c)
+- on JarvisLabs.ai use this [direct link](https://jarvislabs.ai/templates/axolotl)
+- on RunPod use this [direct link](https://runpod.io/gsc?template=v2ickqhz9s&ref=6i7fkpdz)
+
+#### Bare Metal Cloud GPU
+
+##### LambdaLabs
+
+  <details>
+
+  <summary>Click to Expand</summary>
+
+  1. Install python
+  ```bash
+  sudo apt update
+  sudo apt install -y python3.10
+
+  sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.10 1
+  sudo update-alternatives --config python # pick 3.10 if given option
+  python -V # should be 3.10
+
+  ```
+
+  2. Install pip
+  ```bash
+  wget https://bootstrap.pypa.io/get-pip.py
+  python get-pip.py
+  ```
+
+  3. Install Pytorch https://pytorch.org/get-started/locally/
+
+  4. Follow instructions on quickstart.
+
+  5. Run
+  ```bash
+  pip3 install protobuf==3.20.3
+  pip3 install -U --ignore-installed requests Pillow psutil scipy
+  ```
+
+  6. Set path
+  ```bash
+  export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH
+  ```
+  </details>
+
+##### GCP
+
+<details>
+
+<summary>Click to Expand</summary>
+
+Use a Deeplearning linux OS with cuda and pytorch installed. Then follow instructions on quickstart.
+
+Make sure to run the below to uninstall xla.
+```bash
+pip uninstall -y torch_xla[tpu]
+```
+
+</details>
+
+#### Windows
+Please use WSL or Docker!
+
+#### Mac
+
+Use the below instead of the install method in QuickStart.
+```
+pip3 install --no-build-isolation -e '.'
+```
+More info: [mac.md](/docs/mac.qmd)
+
+#### Google Colab
+
+Please use this example [notebook](examples/colab-notebooks/colab-axolotl-example.ipynb).
+
+#### Launching on public clouds via SkyPilot
+To launch on GPU instances (both on-demand and spot instances) on 7+ clouds (GCP, AWS, Azure, OCI, and more), you can use [SkyPilot](https://skypilot.readthedocs.io/en/latest/index.html):
+
+```bash
+pip install "skypilot-nightly[gcp,aws,azure,oci,lambda,kubernetes,ibm,scp]"  # choose your clouds
+sky check
+```
+
+Get the [example YAMLs](https://github.com/skypilot-org/skypilot/tree/master/llm/axolotl) of using Axolotl to finetune `mistralai/Mistral-7B-v0.1`:
+```
+git clone https://github.com/skypilot-org/skypilot.git
+cd skypilot/llm/axolotl
+```
+
+Use one command to launch:
+```bash
+# On-demand
+HF_TOKEN=xx sky launch axolotl.yaml --env HF_TOKEN
+
+# Managed spot (auto-recovery on preemption)
+HF_TOKEN=xx BUCKET=<unique-name> sky spot launch axolotl-spot.yaml --env HF_TOKEN --env BUCKET
+```
+
+#### Launching on public clouds via dstack
+To launch on GPU instance (both on-demand and spot instances) on public clouds (GCP, AWS, Azure, Lambda Labs, TensorDock, Vast.ai, and CUDO), you can use [dstack](https://dstack.ai/).
+
+Write a job description in YAML as below:
+
+```yaml
+# dstack.yaml
+type: task
+
+image: axolotlai/axolotl-cloud:main-latest
+
+env:
+  - HUGGING_FACE_HUB_TOKEN
+  - WANDB_API_KEY
+
+commands:
+  - accelerate launch -m axolotl.cli.train config.yaml
+
+ports:
+  - 6006
+
+resources:
+  gpu:
+    memory: 24GB..
+    count: 2
+```
+
+then, simply run the job with `dstack run` command. Append `--spot` option if you want spot instance. `dstack run` command will show you the instance with cheapest price across multi cloud services:
+
+```bash
+pip install dstack
+HUGGING_FACE_HUB_TOKEN=xxx WANDB_API_KEY=xxx dstack run . -f dstack.yaml # --spot
+```
+
+For further and fine-grained use cases, please refer to the official [dstack documents](https://dstack.ai/docs/) and the detailed description of [axolotl example](https://github.com/dstackai/dstack/tree/master/examples/fine-tuning/axolotl) on the official repository.
+
+### Dataset
+
+Axolotl supports a variety of dataset formats.  It is recommended to use a JSONL.  The schema of the JSONL depends upon the task and the prompt template you wish to use.  Instead of a JSONL, you can also use a HuggingFace dataset with columns for each JSONL field.
+
+See [the documentation](https://axolotl-ai-cloud.github.io/axolotl/docs/dataset-formats/) for more information on how to use different dataset formats.
+
+### Config
+
+See [examples](examples) for quick start. It is recommended to duplicate and modify to your needs. The most important options are:
+
+- model
+  ```yaml
+  base_model: ./llama-7b-hf # local or huggingface repo
+  ```
+  Note: The code will load the right architecture.
+
+- dataset
+  ```yaml
+  datasets:
+      # huggingface repo
+    - path: vicgalle/alpaca-gpt4
+      type: alpaca
+
+      # huggingface repo with specific configuration/subset
+    - path: EleutherAI/pile
+      name: enron_emails
+      type: completion # format from earlier
+      field: text # Optional[str] default: text, field to use for completion data
+
+      # huggingface repo with multiple named configurations/subsets
+    - path: bigcode/commitpackft
+      name:
+        - ruby
+        - python
+        - typescript
+      type: ... # unimplemented custom format
+
+      # chat_template https://axolotl-ai-cloud.github.io/axolotl/docs/dataset-formats/conversation.html#chat_template
+    - path: ...
+      type: chat_template
+      chat_template: chatml # defaults to tokenizer's chat_template
+
+      # local
+    - path: data.jsonl # or json
+      ds_type: json # see other options below
+      type: alpaca
+
+      # dataset with splits, but no train split
+    - path: knowrohit07/know_sql
+      type: context_qa.load_v2
+      train_on_split: validation
+
+      # loading from s3 or gcs
+      # s3 creds will be loaded from the system default and gcs only supports public access
+    - path: s3://path_to_ds # Accepts folder with arrow/parquet or file path like above. Supports s3, gcs.
+      ...
+
+      # Loading Data From a Public URL
+      # - The file format is `json` (which includes `jsonl`) by default. For different formats, adjust the `ds_type` option accordingly.
+    - path: https://some.url.com/yourdata.jsonl # The URL should be a direct link to the file you wish to load. URLs must use HTTPS protocol, not HTTP.
+      ds_type: json # this is the default, see other options below.
+  ```
+
+- loading
+  ```yaml
+  load_in_4bit: true
+  load_in_8bit: true
+
+  bf16: auto # require >=ampere, auto will detect if your GPU supports this and choose automatically.
+  fp16: # leave empty to use fp16 when bf16 is 'auto'. set to false if you want to fallback to fp32
+  tf32: true # require >=ampere
+
+  bfloat16: true # require >=ampere, use instead of bf16 when you don't want AMP (automatic mixed precision)
+  float16: true # use instead of fp16 when you don't want AMP
+  ```
+  Note: Repo does not do 4-bit quantization.
+
+- lora
+  ```yaml
+  adapter: lora # 'qlora' or leave blank for full finetune
+  lora_r: 8
+  lora_alpha: 16
+  lora_dropout: 0.05
+  lora_target_modules:
+    - q_proj
+    - v_proj
+  ```
+
+#### All Config Options
+
+See [these docs](docs/config.qmd) for all config options.
+
+### Train
+
+Run
+```bash
+accelerate launch -m axolotl.cli.train your_config.yml
+```
+
+> [!TIP]
+> You can also reference a config file that is hosted on a public URL, for example `accelerate launch -m axolotl.cli.train https://yourdomain.com/your_config.yml`
+
+#### Preprocess dataset
+
+You can optionally pre-tokenize dataset with the following before finetuning.
+This is recommended for large datasets.
+
+- Set `dataset_prepared_path:` to a local folder for saving and loading pre-tokenized dataset.
+- (Optional): Set `push_dataset_to_hub: hf_user/repo` to push it to Huggingface.
+- (Optional): Use `--debug` to see preprocessed examples.
+
+```bash
+python -m axolotl.cli.preprocess your_config.yml
+```
+
+#### Multi-GPU
+
+Below are the options available in axolotl for training with multiple GPUs. Note that DeepSpeed
+is the recommended multi-GPU option currently because FSDP may experience
+[loss instability](https://github.com/huggingface/transformers/issues/26498).
+
+##### DeepSpeed
+
+Deepspeed is an optimization suite for multi-gpu systems allowing you to train much larger models than you
+might typically be able to fit into your GPU's VRAM. More information about the various optimization types
+for deepspeed is available at https://huggingface.co/docs/accelerate/main/en/usage_guides/deepspeed#what-is-integrated
+
+We provide several default deepspeed JSON configurations for ZeRO stage 1, 2, and 3.
+
+```yaml
+deepspeed: deepspeed_configs/zero1.json
+```
+
+```shell
+accelerate launch -m axolotl.cli.train examples/llama-2/config.yml --deepspeed deepspeed_configs/zero1.json
+```
+
+##### FSDP
+
+- llama FSDP
+```yaml
+fsdp:
+  - full_shard
+  - auto_wrap
+fsdp_config:
+  fsdp_offload_params: true
+  fsdp_state_dict_type: FULL_STATE_DICT
+  fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer
+```
+
+##### FSDP + QLoRA
+
+Axolotl supports training with FSDP and QLoRA, see [these docs](docs/fsdp_qlora.qmd) for more information.
+
+##### Weights & Biases Logging
+
+Make sure your `WANDB_API_KEY` environment variable is set (recommended) or you login to wandb with `wandb login`.
+
+- wandb options
+```yaml
+wandb_mode:
+wandb_project:
+wandb_entity:
+wandb_watch:
+wandb_name:
+wandb_log_model:
+```
+
+##### Comet Logging
+
+Make sure your `COMET_API_KEY` environment variable is set (recommended) or you login to wandb with `comet login`.
+
+- wandb options
+```yaml
+use_comet:
+comet_api_key:
+comet_workspace:
+comet_project_name:
+comet_experiment_key:
+comet_mode:
+comet_online:
+comet_experiment_config:
+```
+
+##### Special Tokens
+
+It is important to have special tokens like delimiters, end-of-sequence, beginning-of-sequence in your tokenizer's vocabulary.  This will help you avoid tokenization issues and help your model train better.  You can do this in axolotl like this:
+
+```yml
+special_tokens:
+  bos_token: "<s>"
+  eos_token: "</s>"
+  unk_token: "<unk>"
+tokens: # these are delimiters
+  - "<|im_start|>"
+  - "<|im_end|>"
+```
+
+When you include these tokens in your axolotl config, axolotl adds these tokens to the tokenizer's vocabulary.
+
+##### Liger Kernel
+
+Liger Kernel: Efficient Triton Kernels for LLM Training
+
+https://github.com/linkedin/Liger-Kernel
+
+Liger (LinkedIn GPU Efficient Runtime) Kernel is a collection of Triton kernels designed specifically for LLM training.
+It can effectively increase multi-GPU training throughput by 20% and reduces memory usage by 60%. The Liger Kernel
+composes well and is compatible with both FSDP and Deepspeed.
+
+```yaml
+plugins:
+  - axolotl.integrations.liger.LigerPlugin
+liger_rope: true
+liger_rms_norm: true
+liger_glu_activation: true
+liger_layer_norm: true
+liger_fused_linear_cross_entropy: true
+```
+
+### Inference Playground
+
+Axolotl allows you to load your model in an interactive terminal playground for quick experimentation.
+The config file is the same config file used for training.
+
+Pass the appropriate flag to the inference command, depending upon what kind of model was trained:
+
+- Pretrained LORA:
+  ```bash
+  python -m axolotl.cli.inference examples/your_config.yml --lora_model_dir="./lora-output-dir"
+  ```
+- Full weights finetune:
+  ```bash
+  python -m axolotl.cli.inference examples/your_config.yml --base_model="./completed-model"
+  ```
+- Full weights finetune w/ a prompt from a text file:
+  ```bash
+  cat /tmp/prompt.txt | python -m axolotl.cli.inference examples/your_config.yml \
+    --base_model="./completed-model" --prompter=None --load_in_8bit=True
+  ```
+-- With gradio hosting
+  ```bash
+  python -m axolotl.cli.inference examples/your_config.yml --gradio
+  ```
+
+Please use `--sample_packing False` if you have it on and receive the error similar to below:
+
+> RuntimeError: stack expects each tensor to be equal size, but got [1, 32, 1, 128] at entry 0 and [1, 32, 8, 128] at entry 1
+
+### Merge LORA to base
+
+The following command will merge your LORA adapater with your base model. You can optionally pass the argument `--lora_model_dir` to specify the directory where your LORA adapter was saved, otherwhise, this will be inferred from `output_dir` in your axolotl config file.  The merged model is saved in the sub-directory `{lora_model_dir}/merged`.
+
+```bash
+python3 -m axolotl.cli.merge_lora your_config.yml --lora_model_dir="./completed-model"
+```
+
+You may need to use the `gpu_memory_limit` and/or `lora_on_cpu` config options to avoid running out of memory. If you still run out of CUDA memory, you can try to merge in system RAM with
+
+```bash
+CUDA_VISIBLE_DEVICES="" python3 -m axolotl.cli.merge_lora ...
+```
+
+although this will be very slow, and using the config options above are recommended instead.
+
+## Common Errors 🧰
+
+See also the [FAQ's](./docs/faq.qmd) and [debugging guide](docs/debugging.qmd).
+
+> If you encounter a 'Cuda out of memory' error, it means your GPU ran out of memory during the training process. Here's how to resolve it:
+
+Please reduce any below
+  - `micro_batch_size`
+  - `eval_batch_size`
+  - `gradient_accumulation_steps`
+  - `sequence_len`
+
+If it does not help, try running without deepspeed and without accelerate (replace "accelerate launch" with "python") in the command.
+
+Using adamw_bnb_8bit might also save you some memory.
+
+> `failed (exitcode: -9)`
+
+Usually means your system has run out of system memory.
+Similarly, you should consider reducing the same settings as when you run out of VRAM.
+Additionally, look into upgrading your system RAM which should be simpler than GPU upgrades.
+
+> RuntimeError: expected scalar type Float but found Half
+
+Try set `fp16: true`
+
+> NotImplementedError: No operator found for `memory_efficient_attention_forward` ...
+
+Try to turn off xformers.
+
+> accelerate config missing
+
+It's safe to ignore it.
+
+> NCCL Timeouts during training
+
+See the [NCCL](docs/nccl.qmd) guide.
+
+
+### Tokenization Mismatch b/w Inference & Training
+
+For many formats, Axolotl constructs prompts by concatenating token ids _after_ tokenizing strings.  The reason for concatenating token ids rather than operating on strings is to maintain precise accounting for attention masks.
+
+If you decode a prompt constructed by axolotl, you might see spaces between tokens (or lack thereof) that you do not expect, especially around delimiters and special tokens.  When you are starting out with a new format, you should always do the following:
+
+1. Materialize some data using `python -m axolotl.cli.preprocess your_config.yml --debug`, and then decode the first few rows with your model's tokenizer.
+2. During inference, right before you pass a tensor of token ids to your model, decode these tokens back into a string.
+3. Make sure the inference string from #2 looks **exactly** like the data you fine tuned on from #1, including spaces and new lines.  If they aren't the same, adjust your inference server accordingly.
+4. As an additional troubleshooting step, you can look at the token ids between 1 and 2 to make sure they are identical.
+
+Having misalignment between your prompts during training and inference can cause models to perform very poorly, so it is worth checking this.  See [this blog post](https://hamel.dev/notes/llm/finetuning/05_tokenizer_gotchas.html) for a concrete example.
+
+## Debugging Axolotl
+
+See [this debugging guide](docs/debugging.qmd) for tips on debugging Axolotl, along with an example configuration for debugging with VSCode.
+
+## Need help? 🙋
+
+Join our [Discord server](https://discord.gg/HhrNrHJPRb) where our community members can help you.
+
+Need dedicated support? Please contact us at [✉️wing@axolotl.ai](ailto:wing@axolotl.ai) for dedicated support options.
--- a/TODO.md
+++ b/TODO.md
@@ -0,0 +1,10 @@
+# todo list
+
+- [] Validation of parameters for combinations that won't work
+
+
+
+## things that are known not to work
+
+- FSDP offload and gradient_checkpointing - https://github.com/pytorch/pytorch/issues/82203
+- adamw_bnb_8bit doesn't play well with FSDP offload
--- a/_quarto.yml
+++ b/_quarto.yml
@@ -1,221 +1,12 @@
 project:
  type: website
-  pre-render: docs/scripts/generate_config_docs.py

-quartodoc:
-  dir: docs/api
-  package: axolotl
-  title: API Reference
-  parser: google
-
-  sections:
-    - title: Core
-      desc: Core functionality for training
-      contents:
-        - train
-        - evaluate
-        - datasets
-        - convert
-        - prompt_tokenizers
-        - logging_config
-        - core.builders.base
-        - core.builders.causal
-        - core.builders.rl
-        - core.training_args
-        - core.chat.messages
-        - core.chat.format.chatml
-        - core.chat.format.llama3x
-        - core.chat.format.shared
-        - core.datasets.chat
-        - core.datasets.transforms.chat_builder
-    - title: CLI
-      desc: Command-line interface
-      contents:
-        - cli.main
-        - cli.train
-        - cli.evaluate
-        - cli.args
-        - cli.art
-        - cli.checks
-        - cli.config
-        - cli.delinearize_llama4
-        - cli.inference
-        - cli.merge_lora
-        - cli.merge_sharded_fsdp_weights
-        - cli.preprocess
-        - cli.quantize
-        - cli.vllm_serve
-        - cli.cloud.base
-        - cli.cloud.modal_
-        - cli.utils
-        - cli.utils.args
-        - cli.utils.fetch
-        - cli.utils.load
-        - cli.utils.sweeps
-        - cli.utils.train
-    - title: Trainers
-      desc: Training implementations
-      contents:
-        - core.trainers.base
-        - core.trainers.trl
-        - core.trainers.mamba
-        - core.trainers.dpo.trainer
-        - core.trainers.grpo.trainer
-        - core.trainers.grpo.sampler
-        - core.trainers.utils
-    - title: Model Loading
-      desc: Functionality for loading and patching models, tokenizers, etc.
-      contents:
-        - loaders.model
-        - loaders.tokenizer
-        - loaders.processor
-        - loaders.adapter
-        - loaders.patch_manager
-        - loaders.constants
-    - title: Mixins
-      desc: Mixin classes for augmenting trainers
-      contents:
-        - core.trainers.mixins.optimizer
-        - core.trainers.mixins.rng_state_loader
-        - core.trainers.mixins.scheduler
-    - title: Context Managers
-      desc: Context managers for altering trainer behaviors
-      contents:
-        - utils.ctx_managers.sequence_parallel
-    - title: Prompt Strategies
-      desc: Prompt formatting strategies
-      contents:
-        - prompt_strategies.base
-        - prompt_strategies.chat_template
-        - prompt_strategies.alpaca_chat
-        - prompt_strategies.alpaca_instruct
-        - prompt_strategies.alpaca_w_system
-        - prompt_strategies.user_defined
-        - prompt_strategies.llama2_chat
-        - prompt_strategies.completion
-        - prompt_strategies.input_output
-        - prompt_strategies.stepwise_supervised
-        - prompt_strategies.metharme
-        - prompt_strategies.orcamini
-        - prompt_strategies.pygmalion
-        - prompt_strategies.messages.chat
-        - prompt_strategies.dpo.chat_template
-        - prompt_strategies.dpo.llama3
-        - prompt_strategies.dpo.chatml
-        - prompt_strategies.dpo.zephyr
-        - prompt_strategies.dpo.user_defined
-        - prompt_strategies.dpo.passthrough
-        - prompt_strategies.kto.llama3
-        - prompt_strategies.kto.chatml
-        - prompt_strategies.kto.user_defined
-        - prompt_strategies.orpo.chat_template
-        - prompt_strategies.bradley_terry.llama3
-    - title: Kernels
-      desc: Low-level performance optimizations
-      contents:
-        - kernels.lora
-        - kernels.geglu
-        - kernels.swiglu
-        - kernels.quantize
-        - kernels.utils
-    - title: Monkey Patches
-      desc: Runtime patches for model optimizations
-      contents:
-        - monkeypatch.llama_attn_hijack_flash
-        - monkeypatch.llama_attn_hijack_xformers
-        - monkeypatch.mistral_attn_hijack_flash
-        - monkeypatch.multipack
-        - monkeypatch.relora
-        - monkeypatch.llama_expand_mask
-        - monkeypatch.lora_kernels
-        - monkeypatch.utils
-        - monkeypatch.btlm_attn_hijack_flash
-        - monkeypatch.llama_patch_multipack
-        - monkeypatch.stablelm_attn_hijack_flash
-        - monkeypatch.trainer_fsdp_optim
-        - monkeypatch.transformers_fa_utils
-        - monkeypatch.unsloth_
-        - monkeypatch.data.batch_dataset_fetcher
-        - monkeypatch.mixtral
-        - monkeypatch.gradient_checkpointing.offload_cpu
-        - monkeypatch.gradient_checkpointing.offload_disk
-    - title: Utils
-      desc: Utility functions
-      contents:
-        - utils.tokenization
-        - utils.chat_templates
-        - utils.lora
-        - utils.model_shard_quant
-        - utils.bench
-        - utils.freeze
-        - utils.trainer
-        - utils.schedulers
-        - utils.distributed
-        - utils.dict
-        - utils.optimizers.adopt
-        - utils.data.pretraining
-        - utils.data.sft
-        - utils.quantization
-    - title: Schemas
-      desc: Pydantic data models for Axolotl config
-      contents:
-        - utils.schemas.config
-        - utils.schemas.model
-        - utils.schemas.training
-        - utils.schemas.datasets
-        - utils.schemas.peft
-        - utils.schemas.trl
-        - utils.schemas.multimodal
-        - utils.schemas.integrations
-        - utils.schemas.enums
-        - utils.schemas.utils
-    - title: Integrations
-      desc: Third-party integrations and extensions
-      contents:
-        - integrations.base
-        - integrations.cut_cross_entropy.args
-        - integrations.grokfast.optimizer
-        - integrations.kd.trainer
-        - integrations.liger.args
-        - integrations.lm_eval.args
-        - integrations.spectrum.args
-    - title: Common
-      desc: Common utilities and shared functionality
-      contents:
-        - common.architectures
-        - common.const
-        - common.datasets
-    - title: Models
-      desc: Custom model implementations
-      contents:
-        - models.mamba.modeling_mamba
-    - title: Data Processing
-      desc: Data processing utilities
-      contents:
-        - utils.collators.core
-        - utils.collators.batching
-        - utils.collators.mamba
-        - utils.collators.mm_chat
-        - utils.samplers.multipack
-    - title: Callbacks
-      desc: Training callbacks
-      contents:
-        - utils.callbacks.perplexity
-        - utils.callbacks.profiler
-        - utils.callbacks.lisa
-        - utils.callbacks.mlflow_
-        - utils.callbacks.comet_
-        - utils.callbacks.qat
 website:
  title: "Axolotl"
-  description: "We make fine-tuning accessible, scalable, and fun"
+  description: "Fine-tuning"
  favicon: favicon.jpg
-
-  google-analytics: "G-9KYCVJBNMQ"
-
  navbar:
-    logo: image/axolotl_logo_digital_white.svg
-    title: false
+    title: Axolotl
    background: dark
    pinned: false
    collapse: false
@@ -234,85 +25,29 @@ website:
      contents:
        - text: Home
          href: index.qmd
-
-        - section: "Getting Started"
+        - section: "How-To Guides"
          contents:
-            - docs/getting-started.qmd
-            - docs/installation.qmd
-            - docs/inference.qmd
-            - docs/cli.qmd
-            - docs/config-reference.qmd
-            - text: "API Reference"
-              href: docs/api
-
+          # TODO Edit folder structure after we have more docs.
+            - docs/debugging.qmd
+            - docs/multipack.qmd
+            - docs/fsdp_qlora.qmd
+            - docs/input_output.qmd
+            - docs/rlhf.qmd
+            - docs/nccl.qmd
+            - docs/mac.qmd
+            - docs/multi-node.qmd
+            - docs/unsloth.qmd
+            - docs/amd_hpc.qmd
        - section: "Dataset Formats"
          contents: docs/dataset-formats/*
-
-        - section: "Deployments"
+        - section: "Reference"
          contents:
-            - docs/docker.qmd
-            - docs/multi-gpu.qmd
-            - docs/multi-node.qmd
-            - docs/ray-integration.qmd
-            - docs/amd_hpc.qmd
-            - docs/mac.qmd
+            - docs/config.qmd
+        - docs/faq.qmd

-        - section: "How To Guides"
-          contents:
-            - docs/multimodal.qmd
-            - docs/rlhf.qmd
-            - docs/reward_modelling.qmd
-            - docs/lr_groups.qmd
-            - docs/lora_optims.qmd
-            - docs/dataset_loading.qmd
-            - docs/qat.qmd
-            - docs/quantize.qmd
-
-        - section: "Core Concepts"
-          contents:
-            - docs/batch_vs_grad.qmd
-            - docs/dataset_preprocessing.qmd
-            - docs/multipack.qmd
-            - docs/mixed_precision.qmd
-            - docs/optimizers.qmd
-
-        - section: "Advanced Features"
-          contents:
-            - docs/fsdp_qlora.qmd
-            - docs/unsloth.qmd
-            - docs/torchao.qmd
-            - docs/custom_integrations.qmd
-            - docs/sequence_parallelism.qmd
-            - docs/gradient_checkpointing.qmd
-            - docs/nd_parallelism.qmd
-
-        - section: "Troubleshooting"
-          contents:
-            - docs/faq.qmd
-            - docs/debugging.qmd
-            - docs/nccl.qmd

 format:
  html:
-    theme: darkly
+    theme: materia
    css: styles.css
    toc: true
-    # Enable better handling of line breaks in markdown
-    preserve-tabs: true
-    html-math-method: mathjax
-    # Improved markdown processing options
-    md-extensions:
-      - markdown_it
-      - def_list
-      - attr_list
-      - fenced_divs
-      - tables
-      - html_admonition
-      - lineblocks
-      - fancy_lists
-    # Control whitespace handling
-    whitespace: preserve
-    # Process newlines in paragraphs
-    wrap: preserve
-    # Better line break handling
-    preserve-linebreaks: true
--- a/cicd/Dockerfile-uv.jinja
+++ b/cicd/Dockerfile-uv.jinja
@@ -1,52 +0,0 @@
-FROM axolotlai/axolotl-base-uv:{{ BASE_TAG }}
-
-ENV TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
-ENV AXOLOTL_EXTRAS="{{ AXOLOTL_EXTRAS }}"
-ENV AXOLOTL_ARGS="{{ AXOLOTL_ARGS }}"
-ENV CUDA="{{ CUDA }}"
-ENV PYTORCH_VERSION="{{ PYTORCH_VERSION }}"
-ENV GITHUB_REF="{{ GITHUB_REF }}"
-ENV GITHUB_SHA="{{ GITHUB_SHA }}"
-ENV NIGHTLY_BUILD="{{ NIGHTLY_BUILD }}"
-ENV HF_HOME="{{ HF_HOME }}"
-
-RUN apt-get update && \
-    apt-get install -y --allow-change-held-packages vim curl nano libnccl2 libnccl-dev ibverbs-providers ibverbs-utils infiniband-diags librdmacm-dev librdmacm1 rdmacm-utils slurm-wlm
-
-WORKDIR /workspace
-
-RUN git clone --depth=1 https://github.com/axolotl-ai-cloud/axolotl.git
-
-WORKDIR /workspace/axolotl
-
-RUN git fetch origin +$GITHUB_REF && \
-    git checkout FETCH_HEAD
-
-# If AXOLOTL_EXTRAS is set, append it in brackets
-RUN if [ "$NIGHTLY_BUILD" = "true" ] ; then \
-        sed -i 's#^transformers.*#transformers @ git+https://github.com/huggingface/transformers.git@main#' requirements.txt; \
-        sed -i 's#^peft.*#peft @ git+https://github.com/huggingface/peft.git@main#' requirements.txt; \
-        sed -i 's#^accelerate.*#accelerate @ git+https://github.com/huggingface/accelerate.git@main#' requirements.txt; \
-        sed -i 's#^trl.*#trl @ git+https://github.com/huggingface/trl.git@main#' requirements.txt; \
-        sed -i 's#^datasets.*#datasets @ git+https://github.com/huggingface/datasets.git@main#' requirements.txt; \
-    fi
-
-RUN uv pip install packaging==23.2 setuptools==75.8.0
-RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
-    else \
-        uv pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
-    fi
-
-RUN python scripts/unsloth_install.py --uv | sh
-RUN python scripts/cutcrossentropy_install.py --uv | sh
-
-# So we can test the Docker image
-RUN uv pip install -r requirements-dev.txt -r requirements-tests.txt
-
-# fix so that git fetch/pull from remote works
-RUN git config remote.origin.fetch "+refs/heads/*:refs/remotes/origin/*" && \
-    git config --get remote.origin.fetch
-
-# helper for huggingface-login cli
-RUN git config --global credential.helper store
--- a/cicd/Dockerfile.jinja
+++ b/cicd/Dockerfile.jinja
@@ -9,10 +9,9 @@ ENV GITHUB_REF="{{ GITHUB_REF }}"
 ENV GITHUB_SHA="{{ GITHUB_SHA }}"
 ENV NIGHTLY_BUILD="{{ NIGHTLY_BUILD }}"
 ENV HF_HOME="{{ HF_HOME }}"
-ENV AXOLOTL_DATASET_PROCESSES="8"

 RUN apt-get update && \
-    apt-get install -y --allow-change-held-packages vim curl nano libnccl2 libnccl-dev ibverbs-providers ibverbs-utils infiniband-diags librdmacm-dev librdmacm1 rdmacm-utils slurm-wlm
+    apt-get install -y --allow-change-held-packages vim curl nano libnccl2 libnccl-dev

 WORKDIR /workspace

@@ -32,11 +31,10 @@ RUN if [ "$NIGHTLY_BUILD" = "true" ] ; then \
        sed -i 's#^datasets.*#datasets @ git+https://github.com/huggingface/datasets.git@main#' requirements.txt; \
    fi

-RUN pip install packaging==23.2 setuptools==75.8.0
 RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
-        pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
+        pip install --no-build-isolation -e .[deepspeed,flash-attn,optimizers,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
    else \
-        pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
+        pip install --no-build-isolation -e .[deepspeed,flash-attn,optimizers] $AXOLOTL_ARGS; \
    fi

 RUN python scripts/unsloth_install.py | sh
--- a/cicd/cicd.sh
+++ b/cicd/cicd.sh
@@ -3,53 +3,7 @@ set -e

 python -c "import torch; assert '$PYTORCH_VERSION' in torch.__version__"

-# Run unit tests with initial coverage report
-pytest -v --durations=10 -n8 \
-  --ignore=tests/e2e/ \
-  --ignore=tests/patched/ \
-  --ignore=tests/cli \
-  /workspace/axolotl/tests/ \
-  --cov=axolotl
-
-# Run lora kernels tests with coverage append
-pytest -v --durations=10 \
-  /workspace/axolotl/tests/e2e/patched/lora_kernels \
-  --cov=axolotl \
-  --cov-append
-
-# Run patched tests excluding lora kernels with coverage append
-pytest --full-trace -vvv --durations=10 \
-  --ignore=tests/e2e/patched/lora_kernels \
-  /workspace/axolotl/tests/e2e/patched \
-  --cov=axolotl \
-  --cov-append
-
-# Run solo tests with coverage append
-pytest -v --durations=10 -n1 \
-  /workspace/axolotl/tests/e2e/solo/ \
-  --cov=axolotl \
-  --cov-append
-
-# Run integration tests with coverage append
-pytest -v --durations=10 \
-  /workspace/axolotl/tests/e2e/integrations/ \
-  --cov=axolotl \
-  --cov-append
-
-pytest -v --durations=10 /workspace/axolotl/tests/cli \
-  --cov=axolotl \
-  --cov-append
-
-# Run remaining e2e tests with coverage append and final report
-pytest -v --durations=10 \
-  --ignore=tests/e2e/solo/ \
-  --ignore=tests/e2e/patched/ \
-  --ignore=tests/e2e/multigpu/ \
-  --ignore=tests/e2e/integrations/ \
-  --ignore=tests/cli \
-  /workspace/axolotl/tests/e2e/ \
-  --cov=axolotl \
-  --cov-append \
-  --cov-report=xml:e2e-coverage.xml
-
-codecov upload-process -t $CODECOV_TOKEN -f e2e-coverage.xml -F e2e,pytorch-${PYTORCH_VERSION} || true
+pytest -v --durations=10 -n8 --ignore=tests/e2e/ --ignore=tests/patched/ /workspace/axolotl/tests/
+pytest -v --durations=10 /workspace/axolotl/tests/e2e/patched/
+pytest -v --durations=10 /workspace/axolotl/tests/e2e/integrations/
+pytest -v --durations=10 --ignore=tests/e2e/patched/ --ignore=tests/e2e/multigpu/ --ignore=tests/e2e/integrations/ /workspace/axolotl/tests/e2e/
--- a/cicd/cleanup.py
+++ b/cicd/cleanup.py
@@ -1,19 +0,0 @@
-"""Modal app to run axolotl GPU cleanup"""
-
-from .single_gpu import VOLUME_CONFIG, app, cicd_image, run_cmd
-
-
-@app.function(
-    image=cicd_image,
-    timeout=60 * 60,
-    cpu=8.0,
-    memory=131072,
-    volumes=VOLUME_CONFIG,
-)
-def cleanup():
-    run_cmd("./cicd/cleanup.sh", "/workspace/axolotl")
-
-
-@app.local_entrypoint()
-def main():
-    cleanup.remote()
--- a/cicd/cleanup.sh
+++ b/cicd/cleanup.sh
@@ -1,6 +0,0 @@
-#!/bin/bash
-set -e
-
-# cleanup old cache files for datasets processing and intermediate mappings
-find /workspace/data/huggingface-cache/hub/datasets -name "cache-*" -type f -mtime +1 -exec rm {} \;
-find /workspace/data/huggingface-cache/hub/datasets -name "*.lock" -type f -mtime +1 -exec rm {} \;
--- a/cicd/e2e_tests.py
+++ b/cicd/e2e_tests.py
@@ -1,20 +0,0 @@
-"""Modal app to run axolotl GPU tests"""
-
-from .single_gpu import GPU_CONFIG, VOLUME_CONFIG, app, cicd_image, run_cmd
-
-
-@app.function(
-    image=cicd_image,
-    gpu=GPU_CONFIG,
-    timeout=120 * 60,  # 90 min
-    cpu=8.0,
-    memory=131072,
-    volumes=VOLUME_CONFIG,
-)
-def cicd_pytest():
-    run_cmd("./cicd/cicd.sh", "/workspace/axolotl")
-
-
-@app.local_entrypoint()
-def main():
-    cicd_pytest.remote()
--- a/cicd/multigpu.py
+++ b/cicd/multigpu.py
@@ -1,7 +1,6 @@
 """
 modal application to run axolotl gpu tests in Modal
 """
-
 # pylint: disable=duplicate-code

 import os
@@ -24,12 +23,11 @@ df_template = template_env.get_template("Dockerfile.jinja")
 df_args = {
    "AXOLOTL_EXTRAS": os.environ.get("AXOLOTL_EXTRAS", ""),
    "AXOLOTL_ARGS": os.environ.get("AXOLOTL_ARGS", ""),
-    "PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.6.0"),
-    "BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.11-cu126-2.6.0"),
-    "CUDA": os.environ.get("CUDA", "126"),
+    "PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.3.1"),
+    "BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.11-cu121-2.3.1"),
+    "CUDA": os.environ.get("CUDA", "121"),
    "GITHUB_REF": os.environ.get("GITHUB_REF", "refs/heads/main"),
    "GITHUB_SHA": os.environ.get("GITHUB_SHA", ""),
-    "CODECOV_TOKEN": os.environ.get("CODECOV_TOKEN", ""),
    "HF_HOME": "/workspace/data/huggingface-cache/hub",
 }

@@ -39,11 +37,15 @@ temp_dir = tempfile.mkdtemp()
 with open(pathlib.Path(temp_dir) / "Dockerfile", "w", encoding="utf-8") as f:
    f.write(dockerfile_contents)

-cicd_image = Image.from_dockerfile(
-    pathlib.Path(temp_dir) / "Dockerfile",
-    force_build=True,
-    gpu="A10G",
-).env(df_args)
+cicd_image = (
+    Image.from_dockerfile(
+        pathlib.Path(temp_dir) / "Dockerfile",
+        force_build=True,
+        gpu="A10G",
+    )
+    .env(df_args)
+    .pip_install("fastapi==0.110.0", "pydantic==2.6.3")
+)

 app = App("Axolotl CI/CD", secrets=[])

@@ -55,7 +57,7 @@ VOLUME_CONFIG = {
 }

 N_GPUS = int(os.environ.get("N_GPUS", 2))
-GPU_CONFIG = f"H100:{N_GPUS}"
+GPU_CONFIG = modal.gpu.H100(count=N_GPUS)


 def run_cmd(cmd: str, run_folder: str):
@@ -69,8 +71,8 @@ def run_cmd(cmd: str, run_folder: str):
@app.function(
    image=cicd_image,
    gpu=GPU_CONFIG,
-    timeout=120 * 60,
-    cpu=16.0,
+    timeout=60 * 60,
+    cpu=8.0,
    memory=131072 * N_GPUS,
    volumes=VOLUME_CONFIG,
 )
--- a/cicd/multigpu.sh
+++ b/cicd/multigpu.sh
@@ -1,25 +1,5 @@
 #!/bin/bash
 set -e

-# Only run two tests at a time to avoid OOM on GPU (with coverage collection)
-pytest -v --durations=10 -n2 \
-  --ignore=/workspace/axolotl/tests/e2e/multigpu/solo/ \
-  --ignore=/workspace/axolotl/tests/e2e/multigpu/patched/ \
-  /workspace/axolotl/tests/e2e/multigpu/ \
-  --cov=axolotl
-
-# Run solo tests with coverage append
-pytest -v --durations=10 -n1 \
-  /workspace/axolotl/tests/e2e/multigpu/solo/ \
-  --cov=axolotl \
-  --cov-append
-
-pytest -v  --durations=10 -n1 /workspace/axolotl/tests/e2e/multigpu/patched/ \
-  --cov=axolotl \
-  --cov-append \
-  --cov-report=xml:multigpu-coverage.xml
-
-# Upload coverage to Codecov if CODECOV_TOKEN is available
-if [ -n "$CODECOV_TOKEN" ]; then
-  codecov upload-process -t "${CODECOV_TOKEN}" -f multigpu-coverage.xml -F multigpu,docker-tests,pytorch-${PYTORCH_VERSION} || true
-fi
+# only run one test at a time so as not to OOM the GPU
+pytest -v -n2 /workspace/axolotl/tests/e2e/multigpu/
--- a/cicd/single_gpu.py
+++ b/cicd/single_gpu.py
@@ -1,5 +1,6 @@
-"""Modal app to run axolotl GPU tests"""
-
+"""
+ modal application to run axolotl gpu tests in Modal
+ """
 # pylint: disable=duplicate-code

 import os
@@ -8,9 +9,8 @@ import tempfile

 import jinja2
 import modal
-import modal.experimental
 from jinja2 import select_autoescape
-from modal import App
+from modal import App, Image

 cicd_path = pathlib.Path(__file__).parent.resolve()

@@ -18,22 +18,18 @@ template_loader = jinja2.FileSystemLoader(searchpath=cicd_path)
 template_env = jinja2.Environment(
    loader=template_loader, autoescape=select_autoescape()
 )
-dockerfile = os.environ.get("E2E_DOCKERFILE", "Dockerfile.jinja")
-df_template = template_env.get_template(dockerfile)
+df_template = template_env.get_template("Dockerfile.jinja")

 df_args = {
    "AXOLOTL_EXTRAS": os.environ.get("AXOLOTL_EXTRAS", ""),
    "AXOLOTL_ARGS": os.environ.get("AXOLOTL_ARGS", ""),
-    "PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.6.0"),
-    "BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.11-cu126-2.6.0"),
-    "CUDA": os.environ.get("CUDA", "126"),
+    "PYTORCH_VERSION": os.environ.get("PYTORCH_VERSION", "2.3.1"),
+    "BASE_TAG": os.environ.get("BASE_TAG", "main-base-py3.11-cu121-2.3.1"),
+    "CUDA": os.environ.get("CUDA", "121"),
    "GITHUB_REF": os.environ.get("GITHUB_REF", "refs/heads/main"),
    "GITHUB_SHA": os.environ.get("GITHUB_SHA", ""),
    "NIGHTLY_BUILD": os.environ.get("NIGHTLY_BUILD", ""),
-    "CODECOV_TOKEN": os.environ.get("CODECOV_TOKEN", ""),
    "HF_HOME": "/workspace/data/huggingface-cache/hub",
-    "PYTHONUNBUFFERED": os.environ.get("PYTHONUNBUFFERED", "1"),
-    "DEEPSPEED_LOG_LEVEL": os.environ.get("DEEPSPEED_LOG_LEVEL", "WARNING"),
 }

 dockerfile_contents = df_template.render(**df_args)
@@ -42,12 +38,16 @@ temp_dir = tempfile.mkdtemp()
 with open(pathlib.Path(temp_dir) / "Dockerfile", "w", encoding="utf-8") as f:
    f.write(dockerfile_contents)

-cicd_image = modal.experimental.raw_dockerfile_image(
-    pathlib.Path(temp_dir) / "Dockerfile",
-    # context_mount=None,
-    force_build=True,
-    # gpu="A10G",
-).env(df_args)
+cicd_image = (
+    Image.from_dockerfile(
+        pathlib.Path(temp_dir) / "Dockerfile",
+        context_mount=None,
+        force_build=True,
+        gpu="A10G",
+    )
+    .env(df_args)
+    .pip_install("fastapi==0.110.0", "pydantic==2.6.3")
+)

 app = App("Axolotl CI/CD", secrets=[])

@@ -59,15 +59,29 @@ VOLUME_CONFIG = {
 }

 N_GPUS = int(os.environ.get("N_GPUS", 1))
-GPU_CONFIG = f"L40S:{N_GPUS}"
+GPU_CONFIG = modal.gpu.A10G(count=N_GPUS)


 def run_cmd(cmd: str, run_folder: str):
    import subprocess  # nosec

-    sp_env = os.environ.copy()
-    sp_env["AXOLOTL_DATASET_PROCESSES"] = "8"
-
    # Propagate errors from subprocess.
-    if exit_code := subprocess.call(cmd.split(), cwd=run_folder, env=sp_env):  # nosec
+    if exit_code := subprocess.call(cmd.split(), cwd=run_folder):  # nosec
        exit(exit_code)  # pylint: disable=consider-using-sys-exit
+
+
+@app.function(
+    image=cicd_image,
+    gpu=GPU_CONFIG,
+    timeout=60 * 60,
+    cpu=8.0,
+    memory=131072,
+    volumes=VOLUME_CONFIG,
+)
+def cicd_pytest():
+    run_cmd("./cicd/cicd.sh", "/workspace/axolotl")
+
+
+@app.local_entrypoint()
+def main():
+    cicd_pytest.remote()
--- a/codecov.yml
+++ b/codecov.yml
@@ -1,57 +0,0 @@
-codecov:
-  require_ci_to_pass: yes
-  notify:
-    wait_for_ci: true
-
-coverage:
-  precision: 2
-  round: down
-  range: "70...100"
-  status:
-    project:
-      default:
-        # basic
-        target: auto
-        threshold: 0%
-        base: auto
-        # advanced
-        branches: null
-        if_no_uploads: error
-        if_not_found: success
-        if_ci_failed: error
-        only_pulls: true
-        flags: null
-        paths: null
-        informational: true
-    patch:
-      default:
-        # basic
-        target: auto
-        threshold: 0%
-        base: auto
-        # advanced
-        branches: null
-        if_no_uploads: error
-        if_not_found: success
-        if_ci_failed: error
-        only_pulls: false
-        flags: null
-        paths: null
-
-parsers:
-  gcov:
-    branch_detection:
-      conditional: yes
-      loop: yes
-      method: no
-      macro: no
-
-comment:
-  layout: "reach,diff,flags,files,footer"
-  behavior: default
-  require_changes: no
-  require_base: no
-  require_head: yes
-
-github_checks:
-  annotations: false
--- a/deepspeed_configs/zero2_torch_compile.json
+++ b/deepspeed_configs/zero2_torch_compile.json
@@ -1,31 +0,0 @@
-{
-  "compile": {
-    "disable": false,
-    "backend": "inductor"
-  },
-  "zero_optimization": {
-    "stage": 2,
-    "offload_optimizer": {
-      "device": "cpu"
-    },
-    "contiguous_gradients": true,
-    "overlap_comm": true
-  },
-  "bf16": {
-    "enabled": "auto"
-  },
-  "fp16": {
-    "enabled": "auto",
-    "auto_cast": false,
-    "loss_scale": 0,
-    "initial_scale_power": 32,
-    "loss_scale_window": 1000,
-    "hysteresis": 2,
-    "min_loss_scale": 1
-  },
-  "gradient_accumulation_steps": "auto",
-  "gradient_clipping": "auto",
-  "train_batch_size": "auto",
-  "train_micro_batch_size_per_gpu": "auto",
-  "wall_clock_breakdown": false
-}
--- a/deepspeed_configs/zero3.json
+++ b/deepspeed_configs/zero3.json
@@ -7,9 +7,9 @@
    "reduce_bucket_size": "auto",
    "stage3_prefetch_bucket_size": "auto",
    "stage3_param_persistence_threshold": "auto",
-    "max_live_parameters": 0,
-    "max_reuse_distance": 0,
-    "gather_16bit_weights_on_model_save": true
+    "stage3_max_live_parameters": 0,
+    "stage3_max_reuse_distance": 0,
+    "stage3_gather_16bit_weights_on_model_save": true
  },
  "bf16": {
    "enabled": "auto"
--- a/deepspeed_configs/zero3_bf16.json
+++ b/deepspeed_configs/zero3_bf16.json
@@ -7,9 +7,9 @@
    "reduce_bucket_size": "auto",
    "stage3_prefetch_bucket_size": "auto",
    "stage3_param_persistence_threshold": "auto",
-    "max_live_parameters": 0,
-    "max_reuse_distance": 0,
-    "gather_16bit_weights_on_model_save": true
+    "stage3_max_live_parameters": 0,
+    "stage3_max_reuse_distance": 0,
+    "stage3_gather_16bit_weights_on_model_save": true
  },
  "bf16": {
    "enabled": true
--- a/deepspeed_configs/zero3_bf16_cpuoffload_all.json
+++ b/deepspeed_configs/zero3_bf16_cpuoffload_all.json
@@ -17,9 +17,9 @@
    "reduce_bucket_size": "auto",
    "stage3_prefetch_bucket_size": "auto",
    "stage3_param_persistence_threshold": "auto",
-    "max_live_parameters": 0,
-    "max_reuse_distance": 0,
-    "gather_16bit_weights_on_model_save": true
+    "stage3_max_live_parameters": 0,
+    "stage3_max_reuse_distance": 0,
+    "stage3_gather_16bit_weights_on_model_save": true
  },
  "bf16": {
    "enabled": true
--- a/deepspeed_configs/zero3_bf16_cpuoffload_params.json
+++ b/deepspeed_configs/zero3_bf16_cpuoffload_params.json
@@ -13,9 +13,9 @@
    "reduce_bucket_size": "auto",
    "stage3_prefetch_bucket_size": "auto",
    "stage3_param_persistence_threshold": "auto",
-    "max_live_parameters": 0,
-    "max_reuse_distance": 0,
-    "gather_16bit_weights_on_model_save": true
+    "stage3_max_live_parameters": 0,
+    "stage3_max_reuse_distance": 0,
+    "stage3_gather_16bit_weights_on_model_save": true
  },
  "bf16": {
    "enabled": true
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -10,9 +10,7 @@ ARG PYTORCH_VERSION="2.1.2"
 ENV PYTORCH_VERSION=$PYTORCH_VERSION

 RUN apt-get update && \
-    apt-get install -y --allow-change-held-packages vim curl nano libnccl2 libnccl-dev rsync s3fs && \
-    rm -rf /var/cache/apt/archives && \
-    rm -rf /var/lib/apt/lists/*
+    apt-get install -y --allow-change-held-packages vim curl nano libnccl2 libnccl-dev rsync s3fs

 WORKDIR /workspace

@@ -22,20 +20,20 @@ WORKDIR /workspace/axolotl

 # If AXOLOTL_EXTRAS is set, append it in brackets
 RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
-        pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
+        pip install --no-build-isolation -e .[deepspeed,flash-attn,optimizers,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
    else \
-        pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray] $AXOLOTL_ARGS; \
-    fi && \
-    python scripts/unsloth_install.py | sh && \
-    python scripts/cutcrossentropy_install.py | sh && \
-    pip install pytest && \
-    pip cache purge
+        pip install --no-build-isolation -e .[deepspeed,flash-attn,optimizers] $AXOLOTL_ARGS; \
+    fi

-# fix so that git fetch/pull from remote works with shallow clone
+RUN python scripts/unsloth_install.py | sh
+RUN python scripts/cutcrossentropy_install.py | sh
+
+# So we can test the Docker image
+RUN pip install pytest
+
+# fix so that git fetch/pull from remote works
 RUN git config remote.origin.fetch "+refs/heads/*:refs/remotes/origin/*" && \
-    git config --get remote.origin.fetch && \
-    git config --global credential.helper store
+    git config --get remote.origin.fetch

-COPY .axolotl-complete.bash /root/.axolotl-complete.bash
-RUN chmod +x /root/.axolotl-complete.bash && \
-    echo 'source /root/.axolotl-complete.bash' >> ~/.bashrc
+# helper for huggingface-login cli
+RUN git config --global credential.helper store
--- a/docker/Dockerfile-base
+++ b/docker/Dockerfile-base
@@ -16,37 +16,24 @@ ENV PYTHON_VERSION=$PYTHON_VERSION
 ENV TORCH_CUDA_ARCH_LIST=$TORCH_CUDA_ARCH_LIST

 RUN apt-get update \
-    && apt-get install -y --no-install-recommends \
-        wget git build-essential ninja-build git-lfs libaio-dev pkg-config \
-        ibverbs-providers ibverbs-utils infiniband-diags  \
-        librdmacm-dev librdmacm1 rdmacm-utils slurm-wlm \
-    && rm -rf /var/cache/apt/archives \
-    && rm -rf /var/lib/apt/lists/* \
+    && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config && rm -rf /var/lib/apt/lists/* \
    && wget \
    https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
    && mkdir /root/.conda \
    && bash Miniconda3-latest-Linux-x86_64.sh -b \
    && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main \
-    && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r \
    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"

 ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"

 WORKDIR /workspace

-RUN python3 -m pip install --upgrade pip && pip3 install -U packaging==23.2 setuptools==75.8.0 wheel && \
-    python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} torchvision --extra-index-url https://download.pytorch.org/whl/cu$CUDA && \
-    CAUSAL_CONV1D_FORCE_CXX11_ABI=TRUE CAUSAL_CONV1D_FORCE_BUILD=TRUE python3 -m pip install --no-cache-dir causal_conv1d==1.5.2 && \
-    python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main" && \
-    python3 -m pip cache purge
+RUN python3 -m pip install --upgrade pip && pip3 install packaging && \
+    python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} --extra-index-url https://download.pytorch.org/whl/cu$CUDA && \
+    python3 -m pip install --no-cache-dir "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" && \
+    python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main"

 RUN git lfs install --skip-repo && \
    pip3 install awscli && \
    # The base image ships with `pydantic==1.8.2` which is not working
-    pip3 install -U --no-cache-dir pydantic==1.10.10 && \
-    pip3 cache purge
-
-RUN if [ "$PYTORCH_VERSION" = "2.6.0" ] && [ "$CUDA" = "124" ] ; then \
-        FLASH_ATTENTION_FORCE_BUILD="TRUE" pip3 install --no-build-isolation flash-attn==2.8.0.post2; \
-    fi
+    pip3 install -U --no-cache-dir pydantic==1.10.10
--- a/docker/Dockerfile-base-next
+++ b/docker/Dockerfile-base-next
@@ -1,38 +0,0 @@
-ARG CUDA_VERSION="12.8.1"
-ARG CUDNN_VERSION="8"
-ARG UBUNTU_VERSION="22.04"
-ARG MAX_JOBS=4
-
-FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
-
-ENV PATH="/root/miniconda3/bin:${PATH}"
-
-ARG PYTHON_VERSION="3.11"
-ARG PYTORCH_VERSION="next"
-ARG CUDA="128"
-ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
-
-ENV PYTHON_VERSION=$PYTHON_VERSION
-ENV TORCH_CUDA_ARCH_LIST=$TORCH_CUDA_ARCH_LIST
-
-RUN apt-get update \
-    && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config && rm -rf /var/lib/apt/lists/* \
-    && wget \
-    https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
-    && mkdir /root/.conda \
-    && bash Miniconda3-latest-Linux-x86_64.sh -b \
-    && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
-
-ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
-
-WORKDIR /workspace
-
-RUN python3 -m pip install --upgrade pip && pip3 install packaging && \
-    python3 -m pip install --no-cache-dir -U torch==2.7.1 --extra-index-url https://download.pytorch.org/whl/test/cu$CUDA && \
-    python3 -m pip install --no-cache-dir "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" && \
-    python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main"
-
-RUN git lfs install --skip-repo && \
-    pip3 install awscli && \
-    pip3 install -U --no-cache-dir pydantic==2.10.6
--- a/docker/Dockerfile-base-nightly
+++ b/docker/Dockerfile-base-nightly
@@ -1,43 +0,0 @@
-ARG CUDA_VERSION="12.8.1"
-ARG CUDNN_VERSION="8"
-ARG UBUNTU_VERSION="22.04"
-ARG MAX_JOBS=4
-
-FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
-
-ENV PATH="/root/miniconda3/bin:${PATH}"
-
-ARG PYTHON_VERSION="3.11"
-ARG PYTORCH_VERSION="nightly"
-ARG CUDA="128"
-ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
-
-ENV PYTHON_VERSION=$PYTHON_VERSION
-ENV TORCH_CUDA_ARCH_LIST=$TORCH_CUDA_ARCH_LIST
-
-RUN apt-get update \
-    && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config && rm -rf /var/lib/apt/lists/* \
-    && wget \
-    https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
-    && mkdir /root/.conda \
-    && bash Miniconda3-latest-Linux-x86_64.sh -b \
-    && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main \
-    && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r \
-    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
-
-ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
-
-WORKDIR /workspace
-
-RUN python3 -m pip install --upgrade pip && pip3 install -U packaging==23.2 setuptools==75.8.0 wheel && \
-    python3 -m pip install --no-cache-dir -U torch --extra-index-url https://download.pytorch.org/whl/nightly/cu$CUDA && \
-    python3 -m pip install --no-cache-dir "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" && \
-    python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main" && \
-    python3 -m pip cache purge
-
-RUN git lfs install --skip-repo && \
-    pip3 install awscli && \
-    # The base image ships with `pydantic==1.8.2` which is not working
-    pip3 install -U --no-cache-dir pydantic==1.10.10 && \
-    pip3 cache purge
--- a/docker/Dockerfile-cloud
+++ b/docker/Dockerfile-cloud
@@ -14,17 +14,13 @@ COPY scripts/motd /etc/motd

 RUN pip install jupyterlab notebook ipywidgets && \
    jupyter lab clean
-RUN apt update && \
-    apt install --yes --no-install-recommends openssh-server tmux iproute2 nvtop && \
-    rm -rf /var/cache/apt/archives && \
-    rm -rf /var/lib/apt/lists/* && \
+RUN apt install --yes --no-install-recommends openssh-server tmux && \
    mkdir -p ~/.ssh && \
    chmod 700 ~/.ssh && \
    printf "\n[[ -z \"\$TMUX\"  ]] && { tmux attach-session -t ssh_tmux || tmux new-session -s ssh_tmux; exit; }\n" >> ~/.bashrc && \
    printf "[ ! -z \"\$TERM\" -a -r /etc/motd ] && cat /etc/motd\n" >> ~/.bashrc && \
    chmod +x /workspace/axolotl/scripts/cloud-entrypoint.sh && \
-    chmod +x /root/cloud-entrypoint.sh && \
-    echo 'set-option -g history-limit 5000' >> ~/.tmux.conf
+    chmod +x /root/cloud-entrypoint.sh

 ENTRYPOINT ["/root/cloud-entrypoint.sh"]
 CMD ["sleep", "infinity"]
--- a/docker/Dockerfile-cloud-no-tmux
+++ b/docker/Dockerfile-cloud-no-tmux
@@ -9,15 +9,13 @@ ENV HF_HUB_ENABLE_HF_TRANSFER="1"
 EXPOSE 8888
 EXPOSE 22

-COPY scripts/cloud-entrypoint.sh /root/cloud-entrypoint.sh
+COPY scripts/cloud-entrypoint-term.sh /root/cloud-entrypoint.sh
 COPY scripts/motd /etc/motd

 RUN pip install jupyterlab notebook ipywidgets && \
    jupyter lab clean
-RUN apt update && \
-    apt install --yes --no-install-recommends openssh-server tmux iproute2 nvtop ibverbs-providers ibverbs-utils infiniband-diags librdmacm-dev librdmacm1 rdmacm-utils slurm-wlm && \
-    rm -rf /var/cache/apt/archives && \
-    rm -rf /var/lib/apt/lists/* && \
+RUN apt install --yes --no-install-recommends openssh-server tmux sudo && \
+    pip3 install -U --no-cache-dir grpcio ray[default]==2.9.3 && \
    mkdir -p ~/.ssh && \
    chmod 700 ~/.ssh && \
    printf "[ ! -z \"\$TERM\" -a -r /etc/motd ] && cat /etc/motd\n" >> ~/.bashrc && \
--- a/docker/Dockerfile-uv-base
+++ b/docker/Dockerfile-uv-base
@@ -1,36 +0,0 @@
-ARG CUDA_VERSION="12.6.3"
-ARG CUDNN_VERSION=""
-ARG UBUNTU_VERSION="22.04"
-ARG MAX_JOBS=4
-
-FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
-
-ARG PYTHON_VERSION="3.11"
-ARG PYTORCH_VERSION="2.6.0"
-ARG CUDA="126"
-ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
-
-ENV PYTHON_VERSION=$PYTHON_VERSION
-ENV TORCH_CUDA_ARCH_LIST=$TORCH_CUDA_ARCH_LIST
-ENV UV_TORCH_BACKEND="cu${CUDA}"
-
-RUN apt-get update \
-    && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config curl && rm -rf /var/lib/apt/lists/* \
-    && git lfs install --skip-repo \
-    && curl -LsSf https://astral.sh/uv/install.sh | sh
-
-ENV PATH="/root/.local/bin:${PATH}"
-
-RUN uv python install ${PYTHON_VERSION}
-
-WORKDIR /workspace
-
-RUN uv venv --no-project --relocatable axolotl-venv
-
-ENV PATH="/workspace/axolotl-venv/bin:${PATH}"
-
-RUN uv pip install packaging setuptools wheel psutil \
-    && uv pip install torch==${PYTORCH_VERSION} \
-    && uv pip install --no-build-isolation "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" \
-    && uv pip install "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main" \
-    && uv pip install awscli pydantic
--- a/docs/.gitignore
+++ b/docs/.gitignore
@@ -1,5 +1,2 @@
 /.quarto/
 _site/
-/api/*.qmd
-/api/*.html
-config-reference.qmd
--- a/docs/amd_hpc.qmd
+++ b/docs/amd_hpc.qmd
@@ -1,5 +1,5 @@
 ---
-title: AMD GPUs on HPC Systems
+title: Training with AMD GPUs on HPC Systems
 description: A comprehensive guide for using Axolotl on distributed systems with AMD GPUs
 ---

--- a/docs/cli.qmd
+++ b/docs/cli.qmd
@@ -1,341 +0,0 @@
---
-title: "Command Line Interface (CLI)"
-format:
-  html:
-    toc: true
-    toc-expand: 2
-    toc-depth: 3
-execute:
-  enabled: false
---
-
-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.
-
-
-## Basic Commands
-
-All Axolotl commands follow this general structure:
-
-```bash
-axolotl <command> [config.yml] [options]
-```
-
-The config file can be local or a URL to a raw YAML file.
-
-### Launcher Arguments
-
-For commands that support multi-GPU (`train`, `evaluate`, ...), you can pass launcher-specific arguments using the `--` separator:
-
-```bash
-# Pass torchrun arguments
-axolotl train config.yml --launcher torchrun -- --nproc_per_node=2 --nnodes=1
-
-# Pass accelerate arguments
-axolotl train config.yml --launcher accelerate -- --config_file=accelerate_config.yml --num_processes=4
-```
-
-Arguments after `--` are passed directly to the launcher (torchrun, accelerate launch, etc.).
-
-## Command Reference
-
-### fetch
-
-Downloads example configurations and deepspeed configs to your local machine.
-
-```bash
-# Get example YAML files
-axolotl fetch examples
-
-# Get deepspeed config files
-axolotl fetch deepspeed_configs
-
-# Specify custom destination
-axolotl fetch examples --dest path/to/folder
-```
-
-### preprocess
-
-Preprocesses and tokenizes your dataset before training. This is recommended for large datasets.
-
-```bash
-# Basic preprocessing
-axolotl preprocess config.yml
-
-# Preprocessing with one GPU
-CUDA_VISIBLE_DEVICES="0" axolotl preprocess config.yml
-
-# Debug mode to see processed examples
-axolotl preprocess config.yml --debug
-
-# Debug with limited examples
-axolotl preprocess config.yml --debug --debug-num-examples 5
-```
-
-Configuration options:
-
-```yaml
-dataset_prepared_path: Local folder for saving preprocessed data
-push_dataset_to_hub: HuggingFace repo to push preprocessed data (optional)
-```
-
-### train
-
-Trains or fine-tunes a model using the configuration specified in your YAML file.
-
-```bash
-# Basic training
-axolotl train config.yml
-
-# Train and set/override specific options
-axolotl train config.yml \
-    --learning-rate 1e-4 \
-    --micro-batch-size 2 \
-    --num-epochs 3
-
-# Training without accelerate
-axolotl train config.yml --launcher python
-
-# Pass launcher-specific arguments using -- separator
-axolotl train config.yml --launcher torchrun -- --nproc_per_node=2 --nnodes=1
-axolotl train config.yml --launcher accelerate -- --config_file=accelerate_config.yml
-
-# Resume training from checkpoint
-axolotl train config.yml --resume-from-checkpoint path/to/checkpoint
-```
-
-It is possible to run sweeps over multiple hyperparameters by passing in a sweeps config.
-
-```bash
-# Basic training with sweeps
-axolotl train config.yml --sweep path/to/sweep.yaml
-```
-
-Example sweep config:
-```yaml
-_:
-  # This section is for dependent variables we need to fix
-  - load_in_8bit: false
-    load_in_4bit: false
-    adapter: lora
-  - load_in_8bit: true
-    load_in_4bit: false
-    adapter: lora
-
-# These are independent variables
-learning_rate: [0.0003, 0.0006]
-lora_r:
-  - 16
-  - 32
-lora_alpha:
-  - 16
-  - 32
-  - 64
-```
-
-
-
-### inference
-
-Runs inference using your trained model in either CLI or Gradio interface mode.
-
-```bash
-# CLI inference with LoRA
-axolotl inference config.yml --lora-model-dir="./outputs/lora-out"
-
-# CLI inference with full model
-axolotl inference config.yml --base-model="./completed-model"
-
-# Gradio web interface
-axolotl inference config.yml --gradio \
-    --lora-model-dir="./outputs/lora-out"
-
-# Inference with input from file
-cat prompt.txt | axolotl inference config.yml \
-    --base-model="./completed-model"
-```
-
-### merge-lora
-
-Merges trained LoRA adapters into the base model.
-
-```bash
-# Basic merge
-axolotl merge-lora config.yml
-
-# Specify LoRA directory (usually used with checkpoints)
-axolotl merge-lora config.yml --lora-model-dir="./lora-output/checkpoint-100"
-
-# Merge using CPU (if out of GPU memory)
-CUDA_VISIBLE_DEVICES="" axolotl merge-lora config.yml
-```
-
-Configuration options:
-
-```yaml
-gpu_memory_limit: Limit GPU memory usage
-lora_on_cpu: Load LoRA weights on CPU
-```
-
-### merge-sharded-fsdp-weights
-
-Merges sharded FSDP model checkpoints into a single combined checkpoint.
-
-```bash
-# Basic merge
-axolotl merge-sharded-fsdp-weights config.yml
-```
-
-### evaluate
-
-Evaluates a model's performance (loss etc) on the train and eval datasets.
-
-```bash
-# Basic evaluation
-axolotl evaluate config.yml
-
-# Evaluation with launcher arguments
-axolotl evaluate config.yml --launcher torchrun -- --nproc_per_node=2
-```
-
-### lm-eval
-
-Runs LM Evaluation Harness on your model.
-
-```bash
-# Basic evaluation
-axolotl lm-eval config.yml
-```
-
-Configuration options:
-
-```yaml
-# List of tasks to evaluate
-lm_eval_tasks:
-  - arc_challenge
-  - hellaswag
-lm_eval_batch_size: # Batch size for evaluation
-output_dir: # Directory to save evaluation results
-```
-
-See [LM Eval Harness](https://github.com/EleutherAI/lm-evaluation-harness) for more details.
-
-### delinearize-llama4
-
-Delinearizes a Llama 4 linearized model into a regular HuggingFace Llama 4 model. This only works with the non-quantized linearized model.
-
-```bash
-axolotl delinearize-llama4 --model path/to/model_dir --output path/to/output_dir
-```
-
-This would be necessary to use with other frameworks. If you have an adapter, merge it with the non-quantized linearized model before delinearizing.
-
-### quantize
-
-Quantizes a model using the quantization configuration specified in your YAML file.
-
-```bash
-axolotl quantize config.yml
-```
-
-See [Quantization](./quantize.qmd) for more details.
-
-
-## Legacy CLI Usage
-
-While the new Click-based CLI is preferred, Axolotl still supports the legacy module-based CLI:
-
-```bash
-# Preprocess
-python -m axolotl.cli.preprocess config.yml
-
-# Train
-accelerate launch -m axolotl.cli.train config.yml
-
-# Inference
-accelerate launch -m axolotl.cli.inference config.yml \
-    --lora_model_dir="./outputs/lora-out"
-
-# Gradio interface
-accelerate launch -m axolotl.cli.inference config.yml \
-    --lora_model_dir="./outputs/lora-out" --gradio
-```
-
-::: {.callout-important}
-When overriding CLI parameters in the legacy CLI, use same notation as in yaml file (e.g., `--lora_model_dir`).
-
-**Note:** This differs from the new Click-based CLI, which uses dash notation (e.g., `--lora-model-dir`). Keep this in mind if you're referencing newer documentation or switching between CLI versions.
-:::
-
-## Remote Compute with Modal Cloud
-
-Axolotl supports running training and inference workloads on Modal cloud infrastructure. This is configured using a
-cloud YAML file alongside your regular Axolotl config.
-
-### Cloud Configuration
-
-Create a cloud config YAML with your Modal settings:
-
-```yaml
-# cloud_config.yml
-provider: modal
-gpu: a100       # Supported: l40s, a100-40gb, a100-80gb, a10g, h100, t4, l4
-gpu_count: 1    # Number of GPUs to use
-timeout: 86400  # Maximum runtime in seconds (24 hours)
-branch: main    # Git branch to use (optional)
-
-volumes:        # Persistent storage volumes
-  - name: axolotl-cache
-    mount: /workspace/cache
-  - name: axolotl-data
-    mount: /workspace/data
-  - name: axolotl-artifacts
-    mount: /workspace/artifacts
-
-secrets:        # Secrets to inject
-  - WANDB_API_KEY
-  - HF_TOKEN
-```
-
-### Running on Modal Cloud
-
-Commands that support the --cloud flag:
-
-```bash
-# Preprocess on cloud
-axolotl preprocess config.yml --cloud cloud_config.yml
-
-# Train on cloud
-axolotl train config.yml --cloud cloud_config.yml
-
-# Run lm-eval on cloud
-axolotl lm-eval config.yml --cloud cloud_config.yml
-```
-
-### Cloud Configuration Options
-
-```yaml
-provider:    # compute provider, currently only `modal` is supported
-gpu:         # GPU type to use
-gpu_count:   # Number of GPUs (default: 1)
-memory:      # RAM in GB (default: 128)
-timeout:     # Maximum runtime in seconds
-timeout_preprocess: # Preprocessing timeout
-branch:      # Git branch to use
-docker_tag:  # Custom Docker image tag
-volumes:     # List of persistent storage volumes
-
-# Environment variables to pass. Can be specified in two ways:
-# 1. As a string: Will load the value from the host computer's environment variables
-# 2. As a key-value pair: Will use the specified value directly
-# Example:
-# env:
-#   - CUSTOM_VAR  # Loads from host's $CUSTOM_VAR
-#   - {CUSTOM_VAR: "value"}  # Uses "value" directly
-env:
-
-# Secrets to inject. Same input format as `env` but for sensitive data.
-secrets:
-  # - HF_TOKEN
-  # - WANDB_API_KEY
-```
--- a/docs/config.qmd
+++ b/docs/config.qmd
@@ -0,0 +1,533 @@
+---
+title: Config options
+description: A complete list of all configuration options.
+---
+
+```yaml
+# This is the huggingface model that contains *.pt, *.safetensors, or *.bin files
+# This can also be a relative path to a model on disk
+base_model: ./llama-7b-hf
+# You can specify an ignore pattern if the model repo contains more than 1 model type (*.pt, etc)
+base_model_ignore_patterns:
+# If the base_model repo on hf hub doesn't include configuration .json files,
+# You can set that here, or leave this empty to default to base_model
+base_model_config: ./llama-7b-hf
+# You can specify to choose a specific model revision from huggingface hub
+revision_of_model:
+# Optional tokenizer configuration path in case you want to use a different tokenizer
+# than the one defined in the base model
+tokenizer_config:
+# If you want to specify the type of model to load, AutoModelForCausalLM is a good choice too
+model_type: AutoModelForCausalLM
+# Corresponding tokenizer for the model AutoTokenizer is a good choice
+tokenizer_type: AutoTokenizer
+# Trust remote code for untrusted source
+trust_remote_code:
+# use_fast option for tokenizer loading from_pretrained, default to True
+tokenizer_use_fast:
+# Whether to use the legacy tokenizer setting, defaults to True
+tokenizer_legacy:
+# Resize the model embeddings when new tokens are added to multiples of 32
+# This is reported to improve training speed on some models
+resize_token_embeddings_to_32x:
+
+# (Internal use only)
+# Used to identify which the model is based on
+is_falcon_derived_model:
+is_llama_derived_model:
+is_qwen_derived_model:
+# Please note that if you set this to true, `padding_side` will be set to "left" by default
+is_mistral_derived_model:
+
+# optional overrides to the base model configuration
+overrides_of_model_config:
+  # RoPE Scaling https://github.com/huggingface/transformers/pull/24653
+  rope_scaling:
+    type: # linear | dynamic
+    factor: # float
+
+# optional overrides to the bnb 4bit quantization configuration
+# https://huggingface.co/docs/transformers/main/main_classes/quantization#transformers.BitsAndBytesConfig
+bnb_config_kwargs:
+  # These are default values
+  llm_int8_has_fp16_weight: false
+  bnb_4bit_quant_type: nf4
+  bnb_4bit_use_double_quant: true
+
+
+# Whether you are training a 4-bit GPTQ quantized model
+gptq: true
+
+# This will attempt to quantize the model down to 8 bits and use adam 8 bit optimizer
+load_in_8bit: true
+# Use bitsandbytes 4 bit
+load_in_4bit:
+
+# Use CUDA bf16
+bf16: true # bool or 'full' for `bf16_full_eval`. require >=ampere
+# Use CUDA fp16
+fp16: true
+# Use CUDA tf32
+tf32: true # require >=ampere
+
+# No AMP (automatic mixed precision)
+bfloat16: true # require >=ampere
+float16: true
+
+# Limit the memory for all available GPUs to this amount (if an integer, expressed in gigabytes); default: unset
+gpu_memory_limit: 20GiB
+# Do the LoRA/PEFT loading on CPU -- this is required if the base model is so large it takes up most or all of the available GPU VRAM, e.g. during a model and LoRA merge
+lora_on_cpu: true
+
+# A list of one or more datasets to finetune the model with
+datasets:
+  # HuggingFace dataset repo | s3://,gs:// path | "json" for local dataset, make sure to fill data_files
+  - path: vicgalle/alpaca-gpt4
+    # The type of prompt to use for training. [alpaca, gpteacher, oasst, reflection]
+    type: alpaca # format | format:<prompt_style> (chat/instruct) | <prompt_strategies>.load_<load_fn>
+    ds_type: # Optional[str] (json|arrow|parquet|text|csv) defines the datatype when path is a file
+    data_files: # Optional[str] path to source data files
+    shards: # Optional[int] number of shards to split data into
+    name: # Optional[str] name of dataset configuration to load
+    train_on_split: train # Optional[str] name of dataset split to load from
+    revision: # Optional[str] The specific revision of the dataset to use when loading from the Hugging Face Hub. This can be a commit hash, tag, or branch name. If not specified, the latest version will be used. This parameter is ignored for local datasets.
+    trust_remote_code: # Optional[bool] Trust remote code for untrusted source
+
+  # Custom user instruction prompt
+  - path: repo
+    type:
+      # The below are defaults. only set what's needed if you use a different column name.
+      system_prompt: ""
+      system_format: "{system}"
+      field_system: system
+      field_instruction: instruction
+      field_input: input
+      field_output: output
+
+      # Customizable to be single line or multi-line
+      # Use {instruction}/{input} as key to be replaced
+      # 'format' can include {input}
+      format: |-
+        User: {instruction} {input}
+        Assistant:
+      # 'no_input_format' cannot include {input}
+      no_input_format: "{instruction} "
+
+      # For `completion` datsets only, uses the provided field instead of `text` column
+      field:
+
+  # Using chat template
+  - path: ...
+    # Set type to `chat_template` to use this strategy
+    type: chat_template
+    # Specify the name of the chat template to use
+    # The name of the chat template to use for training, following values are supported:
+    # - tokenizer_default: Uses the chat template that is available in the tokenizer_config.json. If the chat template is not available in the tokenizer, it will raise an error. This is the default.
+    # - alpaca/inst/chatml/gemma/cohere/llama3/phi_3/deepseek_v2/jamba: These chat templates are available in the axolotl codebase at src/axolotl/utils/chat_templates.py
+    # - tokenizer_default_fallback_*: where * is the name of the chat template to fallback to if the tokenizer does not have a chat template else default to tokenizer. E.g. tokenizer_default_fallback_chatml.
+    # - jinja: Uses a custom jinja template for the chat template. The custom jinja template should be provided in the chat_template_jinja field.
+    chat_template: tokenizer_default
+
+    # Custom jinja chat template. Used only if `chat_template: jinja` or empty.
+    chat_template_jinja:
+
+    # Key containing the messages (default: "messages")
+    field_messages: messages
+    # Key for role in each message (default: "role")
+    message_field_role: role
+    # Key for content in each message (default:  "content")
+    message_field_content: content
+
+    # Optional[Dict[str, List]]. Roles mapping in the messages. The default is:
+    roles:
+      user: ["human", "user"]
+      assistant: ["gpt", "assistant"]
+      system: ["system"]
+      tool: ["tool"]
+
+    # IMPORTANT: The following fields determine which parts of the conversation to train on.
+    # Priority order: message_field_training > message_field_training_detail > train_on_inputs or role in roles_to_train
+    # See examples at `docs/dataset-formats/conversation.qmd`
+    # Note: If the below 4 fields are empty, defaults to training only on the last message.
+
+    # Optional[List[str]]. Roles to train on. The tokens from these roles will be considered for the loss.
+    roles_to_train: ["assistant"]  # default
+    # Optional[str]. Which EOS tokens to train on in the conversation. Possible values are:
+    # - all: train on all EOS tokens
+    # - turn (default): train on the EOS token at the end of each trainable turn
+    # - last: train on the last EOS token in the conversation
+    train_on_eos: last
+    # The key in the message turn that indicates via boolean whether tokens of a turn should be considered for training. Useful to selectively train on certain turns besides the `roles_to_train`.
+    message_field_training: training
+    # The key in the message turn that contains the training details. Useful to selectively train on certain tokens in a turn.
+    # The value of the key is a List[Dict] containing `begin_offset` (start character index in content), `end_offset` (end character index in content), and `train` (boolean whether to train).
+    message_field_training_detail: train_detail
+
+
+# If false, the datasets will not be shuffled and will keep their original order in `datasets`.
+# The same applies to the `test_datasets` option and the `pretraining_dataset` option. Default is true.
+shuffle_merged_datasets: true
+
+Deduplicates datasets and test_datasets with identical entries.
+dataset_exact_deduplication: true
+
+# A list of one or more datasets to eval the model with.
+# You can use either test_datasets, or val_set_size, but not both.
+test_datasets:
+  - path: /workspace/data/eval.jsonl
+    ds_type: json
+    # You need to specify a split. For "json" datasets the default split is called "train".
+    split: train
+    type: completion
+    data_files:
+      - /workspace/data/eval.jsonl
+
+# use RL training: 'dpo', 'ipo', 'kto'
+rl:
+# whether to perform weighting if doing DPO training. Boolean.
+dpo_use_weighting:
+
+# The name of the chat template to use for training, following values are supported:
+# - tokenizer_default: Uses the chat template that is available in the tokenizer_config.json. If the chat template is not available in the tokenizer, it will raise an error. This is the default value.
+# - alpaca/inst/chatml/gemma/cohere/llama3/phi_3/deepseek_v2/jamba: These chat templates are available in the axolotl codebase at src/axolotl/utils/chat_templates.py
+# - tokenizer_default_fallback_*: where * is the name of the chat template to fallback to. E.g. tokenizer_default_fallback_chatml. This is useful when the chat template is not available in the tokenizer.
+# - jinja: Uses a custom jinja template for the chat template. The custom jinja template should be provided in the chat_template_jinja field.
+# The selected chat template will be saved to the tokenizer_config.json for easier inferencing
+# Note: It is recommended to set train_on_inputs to true when using a chat template that is different from the model's default chat template.
+chat_template: tokenizer_default
+# custom jinja template for chat template. This will be only used if chat_template is set to `jinja` or `null` (in which case chat_template is automatically set to `jinja`). Default is null.
+chat_template_jinja: null
+# Changes the default system message
+default_system_message: You are a helpful assistant. Please give a long and detailed answer. # Currently only supports chatml.
+# Axolotl attempts to save the dataset as an arrow after packing the data together so
+# subsequent training attempts load faster, relative path
+dataset_prepared_path: data/last_run_prepared
+# Push prepared dataset to hub
+push_dataset_to_hub: # repo path
+# The maximum number of processes to use while preprocessing your input dataset. This defaults to `os.cpu_count()`
+# if not set.
+dataset_processes: # defaults to os.cpu_count() if not set
+# Keep dataset in memory while preprocessing
+# Only needed if cached dataset is taking too much storage
+dataset_keep_in_memory:
+# push checkpoints to hub
+hub_model_id: # private repo path to push finetuned model
+# how to push checkpoints to hub
+# https://huggingface.co/docs/transformers/v4.31.0/en/main_classes/trainer#transformers.TrainingArguments.hub_strategy
+hub_strategy:
+# Whether to use hf `use_auth_token` for loading datasets. Useful for fetching private datasets
+# Required to be true when used in combination with `push_dataset_to_hub`
+hf_use_auth_token: # boolean
+# How much of the dataset to set aside as evaluation. 1 = 100%, 0.50 = 50%, etc. 0 for no eval.
+val_set_size: 0.04
+# Num shards for whole dataset
+dataset_shard_num:
+# Index of shard to use for whole dataset
+dataset_shard_idx:
+
+# The maximum length of an input to train with, this should typically be less than 2048
+# as most models have a token/context limit of 2048
+sequence_len: 2048
+# Pad inputs so each step uses constant sized buffers
+# This will reduce memory fragmentation and may prevent OOMs, by re-using memory more efficiently
+pad_to_sequence_len:
+# Use efficient multi-packing with block diagonal attention and per sequence position_ids. Recommend set to 'true'
+sample_packing:
+# Set to 'false' if getting errors during eval with sample_packing on.
+eval_sample_packing:
+# You can set these packing optimizations AFTER starting a training at least once.
+# The trainer will provide recommended values for these values.
+sample_packing_eff_est:
+total_num_tokens:
+# Increasing the following values helps with packing, but usually only slightly (<%1.)
+# The number of samples packed at a time.
+sample_packing_group_size: 100000
+# The number of samples which can be packed into one sequence. Increase if using a large sequence_len with many short samples.
+sample_packing_bin_size: 200
+
+# Use batch flattening for speedups when not using sample_packing
+batch_flattening:
+
+# Passed through to transformers when loading the model when launched without accelerate
+# Use `sequential` when training w/ model parallelism to limit memory
+device_map:
+# Defines the max memory usage per gpu on the system. Passed through to transformers when loading the model.
+max_memory:
+
+# If you want to use 'lora' or 'qlora' or leave blank to train all parameters in original model
+adapter: lora
+# If you already have a lora model trained that you want to load, put that here.
+# This means after training, if you want to test the model, you should set this to the value of `output_dir`.
+# Note that if you merge an adapter to the base model, a new subdirectory `merged` will be created under the `output_dir`.
+lora_model_dir:
+
+# LoRA hyperparameters
+# For more details about the following options, see:
+# https://www.anyscale.com/blog/fine-tuning-llms-lora-or-full-parameter-an-in-depth-analysis-with-llama-2
+lora_r: 8
+lora_alpha: 16
+lora_dropout: 0.05
+lora_target_modules:
+  - q_proj
+  - v_proj
+#  - k_proj
+#  - o_proj
+#  - gate_proj
+#  - down_proj
+#  - up_proj
+lora_target_linear: # If true, will target all linear modules
+peft_layers_to_transform: # The layer indices to transform, otherwise, apply to all layers
+
+# If you added new tokens to the tokenizer, you may need to save some LoRA modules because they need to know the new tokens.
+# For LLaMA and Mistral, you need to save `embed_tokens` and `lm_head`. It may vary for other models.
+# `embed_tokens` converts tokens to embeddings, and `lm_head` converts embeddings to token probabilities.
+# https://github.com/huggingface/peft/issues/334#issuecomment-1561727994
+lora_modules_to_save:
+#  - embed_tokens
+#  - lm_head
+
+lora_fan_in_fan_out: false
+
+# LoRA+ hyperparameters
+# For more details about the following options, see:
+# https://arxiv.org/abs/2402.12354  and `src/axolotl/core/train_builder.py`
+loraplus_lr_ratio: # loraplus learning rate ratio lr_B / lr_A. Recommended value is 2^4.
+loraplus_lr_embedding: #  loraplus learning rate for lora embedding layers. Default value is 1e-6.
+
+peft:
+  # Configuration options for loftq initialization for LoRA
+  # https://huggingface.co/docs/peft/developer_guides/quantization#loftq-initialization
+  loftq_config:
+    loftq_bits:  # typically 4 bits
+
+# ReLoRA configuration
+# Must use either 'lora' or 'qlora' adapter, and does not support fsdp or deepspeed
+relora_steps: # Number of steps per ReLoRA restart
+relora_warmup_steps: # Number of per-restart warmup steps
+relora_anneal_steps: # Number of anneal steps for each relora cycle
+relora_prune_ratio: # threshold for optimizer magnitude when pruning
+relora_cpu_offload: # True to perform lora weight merges on cpu during restarts, for modest gpu memory savings
+
+# wandb configuration if you're using it
+# Make sure your `WANDB_API_KEY` environment variable is set (recommended) or you login to wandb with `wandb login`.
+wandb_mode: # "offline" to save run metadata locally and not sync to the server, "disabled" to turn off wandb
+wandb_project: # Your wandb project name
+wandb_entity: # A wandb Team name if using a Team
+wandb_watch:
+wandb_name: # Set the name of your wandb run
+wandb_run_id: # Set the ID of your wandb run
+wandb_log_model: # "checkpoint" to log model to wandb Artifacts every `save_steps` or "end" to log only at the end of training
+
+# mlflow configuration if you're using it
+mlflow_tracking_uri: # URI to mlflow
+mlflow_experiment_name: # Your experiment name
+mlflow_run_name: # Your run name
+hf_mlflow_log_artifacts:  # set to true to copy each saved checkpoint on each save to mlflow artifact registry
+
+# Comet configuration if you're using it
+# Make sure your `COMET_API_KEY` environment variable is set (recommended) or you login to Comet with `comet login`.
+# Check out our documentation for more details https://www.comet.com/docs/v2/api-and-sdk/python-sdk/reference/Experiment-Creation/#comet_ml.start
+use_comet: # Enable or disable Comet integration.
+comet_api_key: # API key for Comet. Recommended to set via `comet login`.
+comet_workspace: # Workspace name in Comet. Defaults to the user's default workspace.
+comet_project_name: # Project name in Comet. Defaults to Uncategorized.
+comet_experiment_key: # Identifier for the experiment. Used to append data to an existing experiment or control the key of new experiments. Default to a random key.
+comet_mode: # Create a new experiment ("create") or log to an existing one ("get"). Default ("get_or_create") auto-selects based on configuration.
+comet_online: # Set to True to log data to Comet server, or False for offline storage. Default is True.
+comet_experiment_config: # Dictionary for additional configuration settings, see the doc for more details.
+
+# Where to save the full-finetuned model to
+output_dir: ./completed-model
+
+# Whether to use torch.compile and which backend to use
+# setting to `auto` will enable torch compile when torch>=2.5.1
+torch_compile:  # Optional[Union[Literal["auto"], bool]]
+torch_compile_backend:  # Optional[str]
+
+# Training hyperparameters
+
+# If greater than 1, backpropagation will be skipped and the gradients will be accumulated for the given number of steps.
+gradient_accumulation_steps: 1
+# The number of samples to include in each batch. This is the number of samples sent to each GPU.
+# Batch size per gpu = micro_batch_size * gradient_accumulation_steps
+micro_batch_size: 2
+eval_batch_size:
+num_epochs: 4
+warmup_steps: 100  # cannot use with warmup_ratio
+warmup_ratio: 0.05  # cannot use with warmup_steps
+learning_rate: 0.00003
+lr_quadratic_warmup:
+logging_steps:
+eval_steps: # Leave empty to eval at each epoch, integers for every N steps. decimal for fraction of total steps
+evals_per_epoch: # number of times per epoch to run evals, mutually exclusive with eval_steps
+save_strategy: # Set to `"no"` to skip checkpoint saves
+save_steps: # Leave empty to save at each epoch
+saves_per_epoch: # number of times per epoch to save a checkpoint, mutually exclusive with save_steps
+save_total_limit: # Checkpoints saved at a time
+# Maximum number of iterations to train for. It precedes num_epochs which means that
+# if both are set, num_epochs will not be guaranteed.
+# e.g., when 1 epoch is 1000 steps => `num_epochs: 2` and `max_steps: 100` will train for 100 steps
+max_steps:
+
+eval_table_size: # Approximate number of predictions sent to wandb depending on batch size. Enabled above 0. Default is 0
+eval_max_new_tokens: # Total number of tokens generated for predictions sent to wandb. Default is 128
+eval_causal_lm_metrics: # HF evaluate metrics used during evaluation. Default is ["sacrebleu", "comet", "ter", "chrf", "perplexity"]
+
+profiler_steps: # enable the pytorch profiler to capture the first N steps of training to the output_dir.
+                # see https://pytorch.org/blog/understanding-gpu-memory-1/ for more information
+                # snapshots can be visualized @ https://pytorch.org/memory_viz
+
+loss_watchdog_threshold: # High loss value, indicating the learning has broken down (a good estimate is ~2 times the loss at the start of training)
+loss_watchdog_patience: # Number of high-loss steps in a row before the trainer aborts (default: 3)
+
+# Save model as safetensors (require safetensors package)
+save_safetensors:
+
+# Whether to mask out or include the human's prompt from the training labels
+train_on_inputs: false
+# Group similarly sized data to minimize padding.
+# May be slower to start, as it must download and sort the entire dataset.
+# Note that training loss may have an oscillating pattern with this enabled.
+group_by_length: false
+
+# Whether to use gradient checkpointing https://huggingface.co/docs/transformers/v4.18.0/en/performance#gradient-checkpointing
+gradient_checkpointing: false
+# additional kwargs to pass to the trainer for gradient checkpointing
+# gradient_checkpointing_kwargs:
+#   use_reentrant: true
+
+# Stop training after this many evaluation losses have increased in a row
+# https://huggingface.co/transformers/v4.2.2/_modules/transformers/trainer_callback.html#EarlyStoppingCallback
+early_stopping_patience: 3
+
+# Specify a scheduler and kwargs to use with the optimizer
+lr_scheduler: # 'one_cycle' | 'log_sweep' | empty for cosine
+lr_scheduler_kwargs:
+cosine_min_lr_ratio: # decay lr to some percentage of the peak lr, e.g. cosine_min_lr_ratio=0.1 for 10% of peak lr
+cosine_constant_lr_ratio: # freeze lr at some percentage of the step, e.g. cosine_constant_lr_ratio=0.8 means start cosine_min_lr at 80% of training step (https://arxiv.org/pdf/2308.04014.pdf)
+
+# For one_cycle optim
+lr_div_factor: # Learning rate div factor
+
+# Specify optimizer
+# Valid values are driven by the Transformers OptimizerNames class, see:
+# https://github.com/huggingface/transformers/blob/95b374952dc27d8511541d6f5a4e22c9ec11fb24/src/transformers/training_args.py#L134
+#
+# Note that not all optimizers may be available in your environment, ex: 'adamw_anyprecision' is part of
+# torchdistx, 'adamw_bnb_8bit' is part of bnb.optim.Adam8bit, etc. When in doubt, it is recommended to start with the optimizer used
+# in the examples/ for your model and fine-tuning use case.
+#
+# Valid values for 'optimizer' include:
+# - adamw_hf
+# - adamw_torch
+# - adamw_torch_fused
+# - adamw_torch_xla
+# - adamw_apex_fused
+# - adopt_adamw (an EXPERIMENTAL optimizer, only for torch version >= 2.5.1)
+# - adafactor
+# - adamw_anyprecision
+# - sgd
+# - adagrad
+# - adamw_bnb_8bit
+# - lion_8bit
+# - lion_32bit
+# - paged_adamw_32bit
+# - paged_adamw_8bit
+# - paged_lion_32bit
+# - paged_lion_8bit
+# - galore_adamw
+# - galore_adamw_8bit
+# - galore_adafactor
+# - galore_adamw_layerwise
+# - galore_adamw_8bit_layerwise
+# - galore_adafactor_layerwise
+optimizer:
+# Dictionary of arguments to pass to the optimizer
+optim_args:
+# For Galore Optimizers the following optim_args are available
+# rank:  # type: int
+# update_proj_gap  # type: int
+# scale  # type: float
+# proj_type:  # type: str, default = std
+
+# The target modules to optimize, i.e. the module names that you would like to train, right now this is used only for GaLore algorithm
+optim_target_modules:
+# - self_attn  # for llama
+# - mlp
+
+# Specify weight decay
+weight_decay:
+# adamw hyperparams
+adam_beta1:
+adam_beta2:
+adam_epsilon:
+# Gradient clipping max norm
+max_grad_norm:
+
+# Augmentation techniques
+# NEFT https://arxiv.org/abs/2310.05914, set this to a number (paper default is 5) to add noise to embeddings
+# currently only supported on Llama and Mistral
+neftune_noise_alpha:
+
+# Whether to bettertransformers
+flash_optimum:
+# Whether to use xformers attention patch https://github.com/facebookresearch/xformers:
+xformers_attention:
+# Whether to use flash attention patch https://github.com/Dao-AILab/flash-attention:
+flash_attention:
+flash_attn_cross_entropy:  # Whether to use flash-attention cross entropy implementation - advanced use only
+flash_attn_rms_norm:  # Whether to use flash-attention rms norm implementation - advanced use only
+flash_attn_fuse_qkv: # Whether to fuse QKV into a single operation
+flash_attn_fuse_mlp: # Whether to fuse part of the MLP into a single operation
+# Whether to use scaled-dot-product attention
+# https://pytorch.org/docs/stable/generated/torch.nn.functional.scaled_dot_product_attention.html
+sdp_attention:
+# Shifted-sparse attention (only llama) - https://arxiv.org/pdf/2309.12307.pdf
+s2_attention:
+# Resume from a specific checkpoint dir
+resume_from_checkpoint:
+# If resume_from_checkpoint isn't set and you simply want it to start where it left off.
+# Be careful with this being turned on between different models.
+auto_resume_from_checkpoints: false
+
+# Don't mess with this, it's here for accelerate and torchrun
+local_rank:
+
+# Add or change special tokens.
+# If you add tokens here, you don't need to add them to the `tokens` list.
+special_tokens:
+  # bos_token: "<s>"
+  # eos_token: "</s>"
+  # unk_token: "<unk>"
+  # pad_token: "[PAD]"
+
+# Add extra tokens.
+tokens:
+
+# FSDP
+fsdp:
+fsdp_config:
+
+# Deepspeed config path. e.g., deepspeed_configs/zero3.json
+deepspeed:
+
+# Advanced DDP Arguments
+ddp_timeout:
+ddp_bucket_cap_mb:
+ddp_broadcast_buffers:
+
+# Path to torch distx for optim 'adamw_anyprecision'
+torchdistx_path:
+
+# Set to HF dataset for type: 'completion' for streaming instead of pre-tokenize
+pretraining_dataset:
+
+# Debug mode
+debug:
+
+# Seed
+seed:
+
+# Allow overwrite yml config using from cli
+strict:
+```
--- a/docs/custom_integrations.qmd
+++ b/docs/custom_integrations.qmd
@@ -1,121 +0,0 @@
---
-title: Custom Integrations
-toc: true
-toc-depth: 3
---
-
-```{python}
-#| echo: false
-
-import os
-import re
-
-def process_readme(integration_name):
-    try:
-        path = f'../src/axolotl/integrations/{integration_name}/README.md'
-        with open(path, 'r') as f:
-            txt = f.read()
-            # Remove h1 headings
-            txt = re.sub(r'^# .*\n?', '', txt, flags=re.MULTILINE)
-            # Convert h2 to h3
-            txt = re.sub(r'^## ', '### ', txt, flags=re.MULTILINE)
-            return txt
-    except FileNotFoundError:
-        return None
-
-def print_section(name, folder_name):
-    output = f"\n## {name}\n"
-    content = process_readme(folder_name)
-    if content:
-        output += content
-    output += f"\nPlease see reference [here](https://github.com/axolotl-ai-cloud/axolotl/tree/main/src/axolotl/integrations/{folder_name})\n"
-    return output
-```
-
-```{python}
-#| output: asis
-#| echo: false
-
-# Introduction text
-print("""
-Axolotl adds custom features through `integrations`. They are located within the `src/axolotl/integrations` directory.
-
-To enable them, please check the respective documentations.
-""")
-
-# Sections
-sections = [
-    ("Cut Cross Entropy", "cut_cross_entropy"),
-    ("Grokfast", "grokfast"),
-    ("Knowledge Distillation (KD)", "kd"),
-    ("Liger Kernels", "liger"),
-    ("Language Model Evaluation Harness (LM Eval)", "lm_eval"),
-    ("Spectrum", "spectrum"),
-    ("LLMCompressor", "llm_compressor")
-]
-
-for folder_name in os.listdir("../src/axolotl/integrations/"):
-    if folder_name in [path for name, path in sections]:
-        # skip if already in sections
-        continue
-    if os.path.exists(f"../src/axolotl/integrations/{folder_name}/README.md"):
-        # grab the first heading in README.md as the section name
-        with open(f"../src/axolotl/integrations/{folder_name}/README.md", "r") as f:
-            txt = f.read()
-            matches = re.search(r'^# (.*)\n?', txt, flags=re.MULTILINE)
-            if matches:
-                name = matches.group(1)
-            else:
-                continue
-            sections.append((name, folder_name))
-
-# sort sections by name
-sections = sorted(sections, key=lambda x: x[0])
-
-for section_name, folder_name in sections:
-    print(print_section(section_name, folder_name))
-```
-
-## Adding a new integration
-
-Plugins can be used to customize the behavior of the training pipeline through [hooks](https://en.wikipedia.org/wiki/Hooking). See [`axolotl.integrations.BasePlugin`](https://github.com/axolotl-ai-cloud/axolotl/blob/main/src/axolotl/integrations/base.py) for the possible hooks.
-
-To add a new integration, please follow these steps:
-
-1. Create a new folder in the `src/axolotl/integrations` directory.
-2. Add any relevant files (`LICENSE`, `README.md`, `ACKNOWLEDGEMENTS.md`, etc.) to the new folder.
-3. Add `__init__.py` and `args.py` files to the new folder.
-  - `__init__.py` should import the integration and hook into the appropriate functions.
-  - `args.py` should define the arguments for the integration.
-4. (If applicable) Add CPU tests under `tests/integrations` or GPU tests under `tests/e2e/integrations`.
-
-::: {.callout-tip}
-
-See [src/axolotl/integrations/cut_cross_entropy](https://github.com/axolotl-ai-cloud/axolotl/tree/main/src/axolotl/integrations/cut_cross_entropy) for a minimal integration example.
-
-:::
-
-::: {.callout-warning}
-
-If you could not load your integration, please ensure you are pip installing in editable mode.
-
-```bash
-pip install -e .
-```
-
-and correctly spelled the integration name in the config file.
-
-```yaml
-plugins:
-  - axolotl.integrations.your_integration_name.YourIntegrationPlugin
-```
-
-:::
-
-::: {.callout-note}
-
-It is not necessary to place your integration in the `integrations` folder. It can be in any location, so long as it's installed in a package in your python env.
-
-See this repo for an example: [https://github.com/axolotl-ai-cloud/diff-transformer](https://github.com/axolotl-ai-cloud/diff-transformer)
-
-:::
--- a/docs/dataset-formats/conversation.qmd
+++ b/docs/dataset-formats/conversation.qmd
@@ -4,15 +4,27 @@ description: Conversation format for supervised fine-tuning.
 order: 3
 ---

+## sharegpt
+
+IMPORTANT: ShareGPT is deprecated!. Please see `chat_template` section below.
+
+
+## pygmalion
+
+```{.json filename="data.jsonl"}
+{"conversations": [{"role": "...", "value": "..."}]}
+```
+
+
 ## chat_template

 Chat Template strategy uses a jinja2 template that converts a list of messages into a prompt. Support using tokenizer's template, a supported template, or custom jinja2.

 ```{.json filename="data.jsonl"}
-{"messages": [{"role": "...", "content": "..."}, {"role": "...", "content": "..."}, ...]}
+{"conversations": [{"role": "...", "content": "..."}]}
 ```

-See [configs](../config-reference.qmd) for full configs and supported templates.
+See `config.qmd` for full configs and supported templates.

 ### Migrating from sharegpt

@@ -32,9 +44,8 @@ datasets:
    type: chat_template

    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
+    message_field_role: from
+    message_field_content: value

 # new (if setting a new chat_template like chatml, gemma, etc)
 chat_template: chatml
@@ -43,18 +54,15 @@ datasets:
    type: chat_template

    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
+    message_field_role: from
+    message_field_content: value
 ```

 We recommend checking the below examples for other usecases.

 ### Examples

-#### Training on last message
-
-(Legacy) Using the default chat template in the tokenizer_config.json on OpenAI messages format, training on only last message.
+1. Using the default chat template in the tokenizer_config.json on OpenAI messages format, training on only last message.

 ```yaml
 datasets:
@@ -64,13 +72,7 @@ datasets:
    train_on_eos:
 ```

-::: {.callout-tip}
-If you receive an error like "`chat_template` choice is `tokenizer_default` but tokenizer's `chat_template` is null.", it means the tokenizer does not have a default `chat_template`. Follow the examples below instead to set a custom `chat_template`.
-:::
-
-#### Overriding default chat template
-
-Using the `gemma` chat template to override the tokenizer_config.json's chat template on OpenAI messages format, training on all assistant messages.
+2. Using the `gemma` chat template to override the tokenizer_config.json's chat template on OpenAI messages format, training on all assistant messages.

 ```yaml
 chat_template: gemma # this overwrites the tokenizer's chat_template
@@ -80,13 +82,7 @@ datasets:
    roles_to_train: ["assistant"]  # default value
 ```

-::: {.callout-note}
-If you want to use built-in chat_template, use `chat_template: tokenizer_default` (this is set by default).
-:::
-
-#### Using default chat template with fallback
-
-Using the tokenizer_config.json's chat template or `chatml` as fallback if the former's chat template does not exist, on OpenAI messages format, training on all assistant messages.
+3. Using the tokenizer_config.json's chat template or `chatml` as fallback if the former's chat template does not exist, on OpenAI messages format, training on all assistant messages.

 ```yaml
 chat_template: tokenizer_default_fallback_chatml # this overwrites the tokenizer's chat_template
@@ -95,9 +91,7 @@ datasets:
    type: chat_template
 ```

-#### Custom Jinja template
-
-Using a custom jinja template on OpenAI messages format, training on all assistant messages.
+4. Using a custom jinja template on OpenAI messages format, training on all assistant messages.

 ```yaml
 # chat_template: jinja # `jinja` will be implied if the `chat_template_jinja` is set and this field is empty
@@ -108,127 +102,7 @@ datasets:
    type: chat_template
 ```

-::: {.callout-important}
-Please make sure that your `tokenizer.eos_token` is same as EOS (End-of-Sequence) token in template. Otherwise, set `eos_token` under `special_tokens: `.
-:::
-
-#### Using template with different token for EOT and EOS
-
- If you are using a template that has a different EOT (End-of-Turn) token from EOS token or multiple EOT tokens (like Mistral V7 Tekken), set the `eot_tokens: ` config. The handling of EOT tokens follows `train_on_eos: ` which defaults to turn.
-
-```yaml
-eot_tokens:
-  - "[/INST]"
-  # - "[/SYSTEM_PROMPT]"
-
-datasets:
-  - path: ...
-    type: chat_template
-
-    # optional
-    train_on_eot: turn  # defaults read from train_on_eos (which defaults to turn)
-```
-
-::: {.callout-tip}
-See [config documentation](../config-reference.qmd) for detailed explanations of "turn", "last", and "all" options for training on tokens.
-:::
-
-::: {.callout-note}
-Using `eot_tokens` requires each token that exists in `chat_template` to be a single token in the tokenizer. Otherwise, the tokenizer will split the token and cause unexpected behavior.
-
-You can add those tokens as new tokens under `tokens: ` or (recommended) override unused added_tokens via `added_tokens_overrides: `. See [config](../config-reference.qmd) for more details.
-:::
-
- Continuing from the previous example, if you want to train on all EOT token trainable turns but only last EOS token, set `train_on_eos: last`.
-
-```yaml
-eot_tokens:
-  - "[/INST]"
-  # ...
-
-datasets:
-  - path: ...
-    type: chat_template
-
-    train_on_eos: last
-    train_on_eot: turn
-```
-
-::: {.callout-tip}
-If EOS token only appears at the end of a prompt, `train_on_eos: last` is equivalent to `train_on_eos: turn`. Therefore, generally, you can leave them to their defaults and omit them.
-:::
-
-
-#### Using tool use
-
-Instead of passing `tools` via the system prompt, an alternative method would be to have the `tools` in a separate column and loaded via `chat_template` to let the template dynamically build it.
-
-```json
-{
-    "tools": [
-        {
-            "type": "...",
-            "function": {
-                "name": "...",
-                "description": "...",
-                "parameters": {
-                    "type": "...",
-                    "properties": {
-                        // ...
-                    },
-                    "required": ["..."],
-                },
-            },
-        },
-    ],
-    "messages": [
-        // ...
-        {
-            "role": "assistant", // call the function via assistant
-            "tool_calls": [
-                {
-                    "id": "...",  // required only for mistral
-                    "type": "function",
-                    "function": {
-                        "name": "...",
-                        "arguments": {
-                            "...": "...",
-                        }
-                    }
-                }
-            ]
-        },
-        {
-            "role": "tool",
-            "tool_call_id": "...",  // required only for mistral
-            "name": "...",
-            "content": "..."
-        },
-    ],
-}
-```
-
-::: {.callout-note}
-Tools need to follow [JSON schema](https://json-schema.org/learn/getting-started-step-by-step).
-:::
-
-Example config for Llama4:
-```yaml
-chat_template: llama4
-datasets:
-  - path: Nanobit/text-tools-2k-test
-    type: chat_template
-    # field_tools: tools # default is `tools`
-```
-
-::: {.callout-tip}
-Look into the `chat_template` you are using to see if it supports `tools` and what the expected role is for the tool answer. In the example above, the tool answer is expected to be in the `tool` or `ipython` role for `llama4` template.
-:::
-
-
-#### Using fine-grained control over token masking
-
-(Advanced) Using fine-grained control over tokens and turns to train in a conversation
+5. (Advanced) Using fine-grained control over tokens and turns to train in a conversation

 For a data sample that looks like:

@@ -266,57 +140,12 @@ datasets:
    type: chat_template
    chat_template: tokenizer_default
    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
+    message_field_role: from
+    message_field_content: value
    roles_to_train: []
    train_on_eos: turn
    message_field_training: train
    message_field_training_detail: train_detail
 ```

-::: {.callout-tip}
-It is not necessary to set both `message_field_training` and `message_field_training_detail` at once.
-:::
-
-#### Reasoning split
-
-(For Qwen3 template only) Enable reasoning split, where the reasoning is split from the content and passed as a separate field into the template.
-
-```yaml
-datasets:
-  - path: ...
-    type: chat_template
-    chat_template: qwen3
-    split_thinking: true
-```
-
-For example, a content can look like:
-
-```json
-{
-  "content": "<think>Some thinking outputs</think>Output after thinking."
-}
-```
-
-After split, it will look like:
-
-```json
-{
-  "reasoning_content": "Some thinking outputs",
-  "content": "Output after thinking..."
-}
-```
-
-
-## sharegpt
-
-::: {.callout-important}
-ShareGPT is deprecated!. Please see [chat_template](#chat_template) section.
-:::
-
-## pygmalion
-
-```{.json filename="data.jsonl"}
-{"conversations": [{"role": "...", "value": "..."}]}
-```
+Tip: It is not necessary to use both `message_field_training` and `message_field_training_detail` at a time.
--- a/docs/dataset-formats/index.qmd
+++ b/docs/dataset-formats/index.qmd
@@ -1,495 +1,14 @@
 ---
 title: Dataset Formats
-description: Guide to Dataset Formats in Axolotl
-back-to-top-navigation: true
-toc: true
-toc-depth: 5
+description: Supported dataset formats.
+listing:
+  fields: [title, description]
+  type: table
+  sort-ui: false
+  filter-ui: false
+  max-description-length: 250
 ---

+Axolotl supports a variety of dataset formats.  It is recommended to use a JSONL format.  The schema of the JSONL depends upon the task and the prompt template you wish to use. Instead of a JSONL, you can also use a HuggingFace dataset with columns for each JSONL field.

-Axolotl is a training framework that aims to make the process convenient yet flexible to users by simply passing a config yaml file.
-
-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.
-
-Axolotl supports 3 kinds of training methods: pre-training, supervised fine-tuning, and preference-based post-training (e.g. DPO, ORPO, PRMs). Each method has their own dataset format which are described below.
-
-::: {.callout-tip}
-
-This guide will mainly use JSONL as an introduction. Please refer to the [dataset loading docs](../dataset_loading.qmd) to understand how to load datasets from other sources.
-
-For `pretraining_dataset:` specifically, please refer to the [Pre-training section](#pre-training).
-:::
-
-## Pre-training
-
-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 [streaming](https://huggingface.co/docs/datasets/en/stream) to only load batches into memory at a time.
-
-A sample format for a pre-training dataset is as follows:
-
-```json
-{"text": "first row"}
-{"text": "second row"}
-...
-```
-
-It is typically recommended to save your dataset as `.jsonl` due to its flexibility and simplicity.
-
-Axolotl supports loading from a Hugging Face hub repo or from local files.
-
-### Pre-training from Hugging Face hub datasets
-
-As an example, to train using a Hugging Face dataset `hf_org/name`, you can pass the following config:
-
-```yaml
-pretraining_dataset: hf_org/name
-```
-
-### Pre-training from local dataset files
-
-Given a few corpus files: `A.jsonl`, `B.jsonl`, and `C.jsonl`, your config will look like the below:
-
-```yaml
-pretraining_dataset:
-  - path: json
-    data_files:
-      - A.jsonl
-      - B.jsonl
-      - C.jsonl
-```
-
-While we recommend `.jsonl`, you can also use the other formats (`csv`, `parquet`, `arrow`, `SQL`, `Webdataset`) that are supported by [`Dataset.load_dataset`](https://huggingface.co/docs/datasets/loading#local-and-remote-files)
-
-### Pre-training without streaming
-
-On the rare case that the dataset is small and can be loaded entirely into memory, another approach to running pre-training is to use the `completion` format. This would mean that the entire dataset is pre-tokenized instead of on-demand in streaming.
-
-One benefit of this is that the tokenization can be performed separately on a CPU-only machine, and then transferred to a GPU machine for training to save costs.
-
-From Hugging Face:
-
-```yaml
-datasets:
-  - path: hf_org/name
-    type: completion
-```
-
-From local files:
-
-```yaml
-datasets:
-  - path: A.jsonl
-    type: completion
-
-  - path: B.jsonl
-    type: completion
-```
-
-::: {.callout-important}
-For `completion` only, Axolotl would split texts if it exceeds the context length into multiple smaller prompts. If you are interested in having this for `pretraining_dataset` too, please let us know or help make a PR!
-:::
-
-### Pre-training dataset configuration tips
-
-#### Setting max_steps
-
-When using streaming for large datasets, Axolotl does not know in advance how large the dataset is and does not know when to stop.
-
-Therefore, it is necessary to set `max_steps: int` in your config for pre-training to run, so that Axolotl knows when to stop training.
-
-One step is equal to `sequence_len * micro_batch_size * gradient_accumulation_steps * total_num_gpus` tokens.
-
-#### Group_by_length
-
-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.
-
-### Reference
-
-Please see docs [here](pretraining.qmd).
-
-## Supervised fine-tuning (SFT)
-
-Supervised fine-tuning is the process of training models to respond to an instruction or chat input.
-
-As there are a wide variety of dataset formats, Axolotl tries to support a majority of the formats available in public datasets.
-
-Axolotl provides four approaches for loading datasets, however, it's easier to work backwards from the dataset you have available to figure out which approach to use.
-
-A flow chart is as follows:
-
-1. Do you already have the dataset tokenized? If yes, check [Pre-Tokenized Dataset](#pre-tokenized-dataset).
-
-2. Do you want to format the dataset yourself and manually choose each section to mask? If yes, check [Template Free Dataset](#template-free-dataset)
-
-3. Is your dataset in a "conversation" format, containing a `list[messages]`? If yes, check [Conversation Dataset](#conversation-dataset)
-
-4. Is your dataset in an "instruct" format, containing `{ instruction, response }`? If yes, check [Instruction Dataset](#instruction-dataset)
-
-If you went through the flow chart and did not find one that matches, it is recommended to preprocess your dataset into one of the above or create a thread on Github Discussion.
-
-::: {.callout-tip}
-You can mix and match within each approach or across approaches to train a model on a variety of datasets.
-:::
-
-### Pre-Tokenized Dataset
-
-We suggest this approach when you want to bring your own tokenized dataset.
-
-Axolotl expects the dataset to have three keys:
-
- `input_ids`: from tokenizing formatted prompt
- `attention_mask`: for masking padding. If you don't add padding, it would be equal to `len(input_ids) * [1]`
- `labels`: this is the same as `input_ids`, however, if you want to mask certain tokens, you would set those indices to `-100`.
-
-::: {.callout-tip}
-Make sure to add BOS/EOS tokens to your prompt and mask it appropriately.
-:::
-
-A config for this would look like:
-
-```yaml
-datasets:
-  - path: A.jsonl
-    type:
-```
-
-::: {.callout-note}
-`type: ` is empty!
-:::
-
-Reference: [Pre-Tokenized Dataset Documentation](tokenized.qmd).
-
-### Template Free Dataset
-
-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 wouldn't suffice.
-
-In the example below, you could see that there is no proper structure. At the same time, it's very flexible as there are no constraints on how your prompt can look.
-
-```json
-{
-    "segments": [
-        {
-            "label": true,
-            "text": "<s>Hello\n"
-        },
-        {
-            "label": true,
-            "text": "hi there!. "
-        },
-        {
-            "label": false,
-            "text": "goodbye "
-        },
-        {
-            "label": true,
-            "text": "farewell</s>"
-        }
-    ]
-}
-```
-
-Each prompt must be have a key called `segments` which is a list of `{ text, label }`.
-
-```yaml
-datasets:
-  - path: A.jsonl
-    type: input_output
-```
-
-Reference: [Template Free Documentation](template_free.qmd).
-
-### Conversation Dataset
-
-`conversation` messages are a list of messages which usually contain a `role` and `content` key.
-
-::: {.callout-tip}
-Fun fact: Axolotl synonymously refers to "chat" messages as `conversation` messages due to how FastChat initially used this term to build a widely used [fastchat conversation](https://github.com/lm-sys/FastChat/blob/main/fastchat/conversation.py) method for formatting chat messages prior to the creation of `chat_templates`.
-:::
-
-#### What are `chat_templates`?
-
-The current most popular and convenient method for inference is to use `chat_templates` for formatting prompts. Axolotl supports using `chat_templates` for training to ensure that the model performs in the same environment as in inference.
-
-Here's a quick rundown on `chat_template`: A `chat_template` is a Jinja2 template which formats a list of messages into a prompt.
-
-An example of a prompt formatted into a popular template called ChatML can be seen below:
-
-Single prompt (pretty-printed):
-```json
-{
-    "messages": [
-        {
-            "role": "user",
-            "content": "Hi"
-        },
-        {
-            "role": "assistant",
-            "content": "How can I help you?"
-        },
-        {
-            "role": "user",
-            "content": "Can you add 3+5?"
-        },
-        {
-            "role": "assistant",
-            "content": "The answer is 8."
-        }
-    ]
-}
-```
-
-The ChatML template is as follows:
-```jinja2
-{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}
-```
-
-The above prompt formatted into this template will result in:
-
-```
-<|im_start|>user
-Hi<|im_end|>
-<|im_start|>assistant
-How can I help you?<|im_end|>
-<|im_start|>user
-Can you add 3+5?<|im_end|>
-<|im_start|>assistant
-The answer is 8.<|im_end|>
-```
-
-By using delimiters (`<|im_start|>` and `<|im_end|>`), a prompt separates different speakers which helps the model identify which portion belongs to whom.
-
-#### Common Conversation Dataset formats
-
-Older conversation datasets with the following format are colloquially called `sharegpt` datasets.
-
-```json
-{"conversations": [{"from": "...", "value": "..."}]}
-```
-
-Newer conversation datasets usually follow the OpenAI format.
-
-```json
-{"messages": [{"role": "...", "content": "..."}]}
-```
-
-Axolotl supports both as well as allowing customization of any kind of key.
-
-#### Chat Template Usage
-
-To properly use this method, it is important to identify three things:
-
-1. Which `chat_template` would you use?
-
-2. What are the keys in your dataset, and what are the possible roles? For example, in OpenAI format, the keys would be `messages`, `role`, and `content`, respectively, whereas the possible roles are `system`, `user`, and `assistant`.
-
-3. What do you want to mask? For instance, only assistant messages, only last message, or nothing.
-
-##### Choosing a `chat_template`
-
-There are a lot of `chat_templates` out there. Axolotl supports the common ones: [supported chat templates](https://github.com/axolotl-ai-cloud/axolotl/blob/860609392184cf62a7e0ca676658b170e059ce6c/src/axolotl/utils/chat_templates.py#L17). For example, to use ChatML, it would be `chat_template: chatml`.
-
-However, it is also possible to use the already configured template within the tokenizer by specifying `chat_template: tokenizer_default`. If you want a fallback (in case some tokenizer does not have it pre-configured), you can do `chat_template: tokenizer_default_fallback_chatml` to fallback to the ChatML template if a tokenizer template was not found.
-
-One last but powerful approach is to bring your own template. This can be set via:
-
-```yaml
-chat_template_jinja: # your template
-```
-
-##### Setting `chat_template` dataset keys
-
-We currently default to OpenAI format for dataset keys, so if that's your current dataset format, there's nothing to do here.
-
-If your dataset format is different, here are the keys you should check (with their defaults):
-
-```yaml
-datasets:
-    ...
-    field_messages: messages  # this should point to the key containing the list of conversations
-    message_property_mappings:  # this is a mapping from keys in your dataset to keys in chat_template
-      role: role
-      content: content
-```
-
-In some `chat_templates` (e.g. [Gemma](https://huggingface.co/google/gemma-2b-it/blob/main/tokenizer_config.json#L1507)), the roles are hardcoded to `user` and `assistant`. Consequently, you may find it necessary to map the roles in your dataset to these above. We currently have some defaults that should work for common datasets, but if you get a `KeyError`, it would be necessary to add mapping for your roles. Here is an example of how it would look like:
-
-```yaml
-datasets:
-    ...
-    roles:
-      assistant:
-        - gpt
-        - model
-      user:
-        - human
-```
-
-In the example above, all `gpt` and `model` values are converted to `assistant`. All `human` values are converted to `user.`
-
-##### Handling masking
-
-The common use case for `chat_template` is for chat messages, therefore, it is common to mask all non-assistant messages. Assistant messages refer to the bot messages that you want the model to learn on.
-
-To train on all `assistant` messages, you would set the following configs.
-
-```yaml
-datasets:
-    ...
-    roles_to_train: ["assistant"]
-    train_on_eos: "turn"
-```
-
-The `train_on_eos` config means that it would mask all EOS tokens for turns that aren't assistant-turns. The other options are: `all` and `last` to choose which EOS to train on.
-
-Perhaps, you want to train on `assistant` and `narrator` roles, you can simply add `narrator` to the list of `roles_to_train`. You would also need to add it to the mapping of `roles` above.
-
-```yaml
-datasets:
-    ...
-    roles_to_train: ["assistant", "narrator"]
-    roles:
-      assistant:
-        - gpt
-        - model
-      user:
-        - human
-      narrator: ["narrator"]
-```
-
-::: {.callout-tip}
-As chat_templates may use hardcoded EOS/EOT tokens that are different from the tokenizer's EOS, it is highly recommended to set them. For example, `ChatML` uses `<|im_end|>` to end turns.
-
-```yaml
-special_tokens:
-  eos_token: <|im_end|>
-```
-
-:::
-
-##### Applying `chat_template`
-
-Once all the above steps are completed, you could combine all these configs together to form a bespoke configuration for your custom dataset.
-
-```yaml
-datasets:
-  - path: A.jsonl
-    type: chat_template
-
-    # step 1
-    chat_template: chatml
-
-    # step 2
-    field_messages: messages
-    message_property_mappings:
-      role: role
-      content: content
-
-    roles:
-      assistant:
-        - gpt
-        - model
-        - assistant
-      user:
-        - human
-        - user
-
-    # step 3
-    roles_to_train: ["assistant"]
-    train_on_eos: "turn"
-
-special_tokens:
-  eos_token: <|im_end|>
-```
-
-If this config were to be applied to the sample dataset above, the output would look as such (which can be retrieved via `axolotl preprocess config.yaml --debug`):
-
-```
-<|im_start|>(-100, 128256) user(-100, 882)
-(-100, 198) Hi(-100, 13347) <|im_end|>(-100, 128257)
-(-100, 198) <|im_start|>(-100, 128256) assistant(-100, 78191)
-(-100, 198) How(4438, 4438)  can(649, 649)  I(358, 358)  help(1520, 1520)  you(499, 499) ?(30, 30) <|im_end|>(128257, 128257)
-(-100, 198) <|im_start|>(-100, 128256) user(-100, 882)
-(-100, 198) Can(-100, 6854)  you(-100, 499)  add(-100, 923)  (-100, 220) 3(-100, 18) +(-100, 10) 5(-100, 20) ?(-100, 30) <|im_end|>(-100, 128257)
-(-100, 198) <|im_start|>(-100, 128256) assistant(-100, 78191)
-(-100, 198) The(791, 791)  answer(4320, 4320)  is(374, 374)  (220, 220) 8(23, 23) .(13, 13) <|im_end|>(128257, 128257)
-(-100, 198)
-```
-
-The first number refers to the label, the second refers to the `token_id`. For example, `-100` labels appear on non-assistant portions, meaning that they are masked during. For assistant portions, the label is the same as the `token_id`.
-
-::: {.callout-note}
-
-If during `preprocess`, there are a lot of warnings of `Could not find content __ boundary`, please check the FAQ section for [chat_templates](../faq.qmd#chat-templates).
-
-:::
-
-#### Reference
-
-Please see docs [here](conversation.qmd).
-
-### Instruction Dataset
-
-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.
-
-An example is of a common format called Alpaca:
-```json
-{"instruction": "...", "input": "...", "output": "..."}
-```
-
-Using those keys, a prompt can be built based on it.
-```
-Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
-
-### Instruction:
-{instruction}
-
-### Input:
-{input}
-
-### Response:
-{output}
-```
-
-This can be configured as such:
-```yaml
-datasets:
-  - path: A.jsonl
-    type: alpaca
-```
-
-Axolotl supports many kinds of instruction dataset. All of them can be found in the [Instruction Dataset Documentation](inst_tune.qmd) with their respective type and sample row format.
-
-#### Custom Instruct Prompt Format
-
-Due to the myriad possibilities of instruction formats, Axolotl allows customizing your own instruction format without having to dive into the code directly.
-
-In the example below, a sample row is used to output in `mistral_v1` format.
-```json
-{"input": "...", "output": "..."}
-```
-
-```yaml
-datasets:
-  - path: repo
-    type:
-      system_prompt: ""
-
-      field_system:
-      field_instruction: input
-      field_input:
-      field_output: output
-
-      # multi-line example with input
-      format: |-
-        [INST] {instruction} {input} [/INST]
-
-      # single-line example without input
-      no_input_format: "[INST] {instruction} [/INST]"
-```
-
-The config sets that the `field_instruction` is actually named `input`, and the `field_input` is empty as we don't have an `input` in this sample. Generally, `instruction` can be thought as the question to the model, and `input` as the additional information with `output` being the response. It is not necessary to have an `input` nor `system`. 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.
-
-Reference: [Custom Instruct Prompt Format Documentation](inst_tune.qmd#how-to-add-custom-prompt-format).
-
-## Reinforcement Learning from Human Feedback (RLHF)
-
-As there are multiple RLHF methods with their own dataset requirements. Please see [RLHF documentation](../rlhf.qmd) for more detail.
+Below are these various formats organized by task:
--- a/docs/dataset-formats/inst_tune.qmd
+++ b/docs/dataset-formats/inst_tune.qmd
@@ -186,4 +186,4 @@ datasets:
      no_input_format: "[INST] {instruction} [/INST]"
 ```

-See full config options under [here](../config-reference.qmd).
+See full config options under [here](../config.qmd).
--- a/docs/dataset-formats/pretraining.qmd
+++ b/docs/dataset-formats/pretraining.qmd
@@ -19,14 +19,8 @@ For pretraining, there is no prompt template or roles.  The only required field
 Axolotl usually loads the entire dataset into memory. This will be challenging for large datasets. Use the following config to enable streaming:

 ```{.yaml filename="config.yaml"}
-pretraining_dataset:
-  - name:
-    path:
-    split:
-    text_column: # column in dataset with the data, usually `text`
-    type: pretrain
-    trust_remote_code:
-    skip: # number of rows of data to skip over from the beginning
+pretraining_dataset: # hf path only
+...
 ```

 :::
--- a/docs/dataset-formats/stepwise_supervised.qmd
+++ b/docs/dataset-formats/stepwise_supervised.qmd
@@ -1,26 +0,0 @@
---
-title: Stepwise Supervised Format
-description: Format for datasets with stepwise completions and labels
-order: 3
---
-
-## Stepwise Supervised
-
-The stepwise supervised format is designed for chain-of-thought (COT) reasoning
-datasets where each example contains multiple completion steps and a preference label
-for each step.
-
-### Example
-
-Here's a simple example of a stepwise supervised dataset entry:
-
-```json
-{
-  "prompt": "Which number is larger, 9.8 or 9.11?",
-  "completions": [
-    "The fractional part of 9.8 is 0.8, while the fractional part of 9.11 is 0.11.",
-    "Since 0.11 is greater than 0.8, the number 9.11 is larger than 9.8."
-  ],
-  "labels": [true, false]
-}
-```
--- a/docs/dataset-formats/template_free.qmd
+++ b/docs/dataset-formats/template_free.qmd
@@ -1,239 +1,7 @@
 ---
 title: Template-Free
 description: Construct prompts without a template.
-toc: true
-toc-depth: 3
 order: 4
 ---

-## Background {#sec-background}
-
-### Masking Inputs {#masking-inputs}
-
-One of the most popular features of
-[axolotl](https://github.com/axolotl-ai-cloud/axolotl) is
-setting the following configuration value:
-
-
-```yaml
-train_on_inputs: false
-```
-
-If you declare a [dataset formats](https://github.com/axolotl-ai-cloud/axolotl?tab=readme-ov-file#dataset)
-such as `alpaca` or `chatml`, axolotl knows what is an input
-(i.e. human) vs. an output (i.e. the assistant) and masks the input
-labels so that your model can focus on predicting the outputs only.
-
-### You may not want prompt templates {#sec-you-may-not-want-prompt-templates}
-
-However, there are many situations where you don't want to use one of
-these formats or templates. This is because they can:
-
-   Add unnecessary boilerplate to your prompts.
-   Create artifacts like special delimiters `<|im_start|>` that can
-    quickly become footguns if you don't include them correctly at
-    inference time.
-   Enforce a *chat* 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.
-   Limit you to only certain roles that the template allows.
-
-### The `input_output` format {#sec-the-inputoutput-format}
-
-You can construct your prompts without a template by using the
-`input_output` format, by setting `type: input_output` in your
-configuration file like this:
-
-**config.yml**
-
-```yaml
-train_on_inputs: false # Mask segments of your data
-datasets:
-  - path: output.jsonl
-    type: input_output  # use template free prompt construction
-```
-
-Unlike `type: completion`, which is also template-free,
-`type: input_output` allows you to mask segments of your text. More
-details on how this works are described below.
-
-## Usage {#sec-usage}
-
-This is how you can use the `input_output` format:
-
-### 1. Prepare Data {#sec-1-prepare-data}
-
-To use the `input_output` format, collect your data in the following
-format into a jsonl file (below is the first row from the file
-`output`.jsonl` pretty printed):
-
-```bash
-$ head -n1 output.jsonl | python -m json.tool
-```
-
-:::{.cell-output .cell-output-stdout}
-    {
-        "segments": [
-            {
-                "label": true,
-                "text": "<s>Hello\n"
-            },
-            {
-                "label": true,
-                "text": "hi there!. "
-            },
-            {
-                "label": false,
-                "text": "goodbye "
-            },
-            {
-                "label": true,
-                "text": "farewell</s>"
-            }
-        ]
-    }
-:::
-
-Set `label:false` when you want to mask a segment of text so that the
-model isn't trained on it. Some things to keep in mind:
-
-> [!IMPORTANT]
-> 1.  **EOS, BOS, spaces, newlines etc. are entirely up to you. Axolotl
-    concatenates all the segments as-is.** The tokenizer doesn't add
-    anything additional. Notice how I added spaces, newlines, `<s>`
-    (BOS), and `</s>` (EOS) myself.
-> 2.  Make sure you check the materialized output to validate that the
-    prompt is getting assembled how you like.
-
-### 2. Use `type: input_output` {#sec-2-use-type-inputoutput}
-
-Let's materialize data with our `output.jsonl` file by setting
-`type: input_output` in our axolotl config:
-
-```yaml
-# training_config.yaml
-base_model: mistralai/Mistral-7B-v0.1
-data_seed: 49
-seed: 49
-
-datasets:
-  - path: output.jsonl
-    type: input_output
-val_set_size: 0.1
-
-sequence_len: 896
-sample_packing: false
-
-micro_batch_size: 2
-gradient_accumulation_steps: 3
-eval_batch_size: 2
-num_epochs: 1
-learning_rate: 0.0002
-
-train_on_inputs: false
-special_tokens:
-  bos_token: "<s>"
-  eos_token: "</s>"
-  unk_token: "<unk>"
-```
-
-You can use the following command to materialize your data. The
-`--debug` flag will print the tokens, along with the labels so you can
-verify that the correct items are being ignored:
-
-```bash
-axolotl preprocess training_config.yaml --debug
-
-...
-[2024-03-05 23:36:46,969] [INFO] [axolotl.check_example_labels:35] [PID:607731] [RANK:0] <s>(1, 1) Hello(22557, 22557)
-(13, 13) hi(12014, 12014) there(736, 736) !(28808, 28808) .(28723, 28723) (28705, 28705) good(-100, 1179) bye(-100, 17664) (-100, 28705) fare(19111, 19111) well(5458, 5458) </s>(2, 2)
-
-```
-
-The format is `decoded_token`(`label`, `token_id`), for example,
-`<s>(1, 1)` means that the token is `<s>`, the label is `1` and the
-token_id is `1`. When the label is `-100` then that token is ignored for
-training.
-
-### 3. Check the prompts {#sec-3-check-the-prompts}
-
-Here is another way to check the materialized output:
-
-```python
-from transformers import AutoTokenizer
-from datasets import load_from_disk
-import yaml
-
-directory = !ls last_run_prepared/
-with open('training_config.yaml', 'r') as f:
-    cfg = yaml.safe_load(f)
-model_id = cfg['base_model']
-tok = AutoTokenizer.from_pretrained(model_id)
-ds = load_from_disk(f'last_run_prepared/{directory[0]}/')
-```
-
-```python
->>> row = ds[0]
->>> print(tok.decode(row['input_ids']))
-<s> Hello
-    hi there!.  goodbye  farewell</s>
-```
-
-We can check that the right tokens are ignored by comparing the labels
-to each token:
-
-```python
-import pandas as pd
-pd.DataFrame([{'token': tok.decode(i), 'label': l, 'id':i} for i,l in
-              zip(row['input_ids'], row['labels'])])
-```
-
-| token | label | id    |
-|-------|-------|-------|
-| 0     | \<s\> | 1     |
-| 1     | Hello | 22557 |
-| 2     | \\n   | 13    |
-| 3     | hi    | 12014 |
-| 4     | there | 736   |
-| 5     | !     | 28808 |
-| 6     | .     | 28723 |
-| 7     |       | 28705 |
-| 8     | good  | -100  |
-| 9     | bye   | -100  |
-| 10    |       | -100  |
-| 11    | fare  | 19111 |
-| 12    | well  | 5458  |
-| 13    | \</s\>| 2     |
-
-
-
-If we look at the input data, the above table seems correct! (The jsonl
-version is repeated below for reference):
-
-
-```bash
-$ head -n1 output.jsonl | python -m json.tool
-```
-
-:::{.cell-output .cell-output-stdout}
-    {
-        "segments": [
-            {
-                "label": true,
-                "text": "<s>Hello\n"
-            },
-            {
-                "label": true,
-                "text": "hi there!. "
-            },
-            {
-                "label": false,
-                "text": "goodbye "
-            },
-            {
-                "label": true,
-                "text": "farewell</s>"
-            }
-        ]
-    }
-:::
+See [these docs](../input_output.qmd).
--- a/docs/dataset_loading.qmd
+++ b/docs/dataset_loading.qmd
@@ -1,268 +0,0 @@
---
-title: Dataset Loading
-description: Understanding how to load datasets from different sources
-back-to-top-navigation: true
-toc: true
-toc-depth: 5
---
-
-## Overview
-
-Datasets can be loaded in a number of different ways depending on the how it is saved (the extension of the file) and where it is stored.
-
-## Loading Datasets
-
-We use the `datasets` library to load datasets and a mix of `load_dataset` and `load_from_disk` to load them.
-
-You may recognize the similar named configs between `load_dataset` and the `datasets` section of the config file.
-
-```yaml
-datasets:
-  - path:
-    name:
-    data_files:
-    split:
-    revision:
-    trust_remote_code:
-```
-
-::: {.callout-tip}
-
-Do not feel overwhelmed by the number of options here. A lot of them are optional. In fact, the most common config to use would be `path` and sometimes `data_files`.
-
-:::
-
-This matches the API of [`datasets.load_dataset`](https://github.com/huggingface/datasets/blob/0b5998ac62f08e358f8dcc17ec6e2f2a5e9450b6/src/datasets/load.py#L1838-L1858), so if you're familiar with that, you will feel right at home.
-
-For HuggingFace's guide to load different dataset types, see [here](https://huggingface.co/docs/datasets/loading).
-
-For full details on the config, see [config-reference.qmd](config-reference.qmd).
-
-::: {.callout-note}
-
-You can set multiple datasets in the config file by more than one entry under `datasets`.
-
-```yaml
-datasets:
-  - path: /path/to/your/dataset
-  - path: /path/to/your/other/dataset
-```
-
-:::
-
-### Local dataset
-
-#### Files
-
-To load a JSON file, you would do something like this:
-
-```python
-from datasets import load_dataset
-
-dataset = load_dataset("json", data_files="data.json")
-```
-
-Which translates to the following config:
-
-```yaml
-datasets:
-  - path: data.json
-    ds_type: json
-```
-
-In the example above, it can be seen that we can just point the `path` to the file or directory along with the `ds_type` to load the dataset.
-
-This works for CSV, JSON, Parquet, and Arrow files.
-
-::: {.callout-tip}
-
-If `path` points to a file and `ds_type` is not specified, we will automatically infer the dataset type from the file extension, so you could omit `ds_type` if you'd like.
-
-:::
-
-#### Directory
-
-If you're loading a directory, you can point the `path` to the directory.
-
-Then, you have two options:
-
-##### Loading entire directory
-
-You do not need any additional configs.
-
-We will attempt to load in the following order:
- datasets saved with `datasets.save_to_disk`
- loading entire directory of files (such as with parquet/arrow files)
-
-```yaml
-datasets:
-  - path: /path/to/your/directory
-```
-
-##### Loading specific files in directory
-
-Provide `data_files` with a list of files to load.
-
-```yaml
-datasets:
-    # single file
-  - path: /path/to/your/directory
-    ds_type: csv
-    data_files: file1.csv
-
-    # multiple files
-  - path: /path/to/your/directory
-    ds_type: json
-    data_files:
-      - file1.jsonl
-      - file2.jsonl
-
-    # multiple files for parquet
-  - path: /path/to/your/directory
-    ds_type: parquet
-    data_files:
-      - file1.parquet
-      - file2.parquet
-
-```
-
-### HuggingFace Hub
-
-The method you use to load the dataset depends on how the dataset was created, whether a folder was uploaded directly or a HuggingFace Dataset was pushed.
-
-::: {.callout-note}
-
-If you're using a private dataset, you will need to enable the `hf_use_auth_token` flag in the root-level of the config file.
-
-:::
-
-#### Folder uploaded
-
-This would mean that the dataset is a single file or file(s) uploaded to the Hub.
-
-```yaml
-datasets:
-  - path: org/dataset-name
-    data_files:
-      - file1.jsonl
-      - file2.jsonl
-```
-
-#### HuggingFace Dataset
-
-This means that the dataset is created as a HuggingFace Dataset and pushed to the Hub via `datasets.push_to_hub`.
-
-```yaml
-datasets:
-  - path: org/dataset-name
-```
-
-::: {.callout-note}
-
-There are some other configs which may be required like `name`, `split`, `revision`, `trust_remote_code`, etc depending on the dataset.
-
-:::
-
-### Remote Filesystems
-
-Via the `storage_options` config under `load_dataset`, you can load datasets from remote filesystems like S3, GCS, Azure, and OCI.
-
-::: {.callout-warning}
-
-This is currently experimental. Please let us know if you run into any issues!
-
-:::
-
-The only difference between the providers is that you need to prepend the path with the respective protocols.
-
-```yaml
-datasets:
-    # Single file
-  - path: s3://bucket-name/path/to/your/file.jsonl
-
-    # Directory
-  - path: s3://bucket-name/path/to/your/directory
-```
-
-For directory, we load via `load_from_disk`.
-
-#### S3
-
-Prepend the path with `s3://`.
-
-The credentials are pulled in the following order:
-
- `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, and `AWS_SESSION_TOKEN` environment variables
- from the `~/.aws/credentials` file
- for nodes on EC2, the IAM metadata provider
-
-::: {.callout-note}
-
-We assume you have credentials setup and not using anonymous access. If you want to use anonymous access, let us know! We may have to open a config option for this.
-
-:::
-
-Other environment variables that can be set can be found in [boto3 docs](https://boto3.amazonaws.com/v1/documentation/api/latest/guide/configuration.html#using-environment-variables)
-
-#### GCS
-
-Prepend the path with `gs://` or `gcs://`.
-
-The credentials are loaded in the following order:
-
- gcloud credentials
- for nodes on GCP, the google metadata service
- anonymous access
-
-#### Azure
-
-##### Gen 1
-
-Prepend the path with `adl://`.
-
-Ensure you have the following environment variables set:
-
- `AZURE_STORAGE_TENANT_ID`
- `AZURE_STORAGE_CLIENT_ID`
- `AZURE_STORAGE_CLIENT_SECRET`
-
-##### Gen 2
-
-Prepend the path with `abfs://` or `az://`.
-
-Ensure you have the following environment variables set:
-
- `AZURE_STORAGE_ACCOUNT_NAME`
- `AZURE_STORAGE_ACCOUNT_KEY`
-
-Other environment variables that can be set can be found in [adlfs docs](https://github.com/fsspec/adlfs?tab=readme-ov-file#setting-credentials)
-
-#### OCI
-
-Prepend the path with `oci://`.
-
-It would attempt to read in the following order:
-
- `OCIFS_IAM_TYPE`, `OCIFS_CONFIG_LOCATION`, and `OCIFS_CONFIG_PROFILE` environment variables
- when on OCI resource, resource principal
-
-Other environment variables:
-
- `OCI_REGION_METADATA`
-
-Please see the [ocifs docs](https://ocifs.readthedocs.io/en/latest/getting-connected.html#Using-Environment-Variables).
-
-### HTTPS
-
-The path should start with `https://`.
-
-```yaml
-datasets:
-  - path: https://path/to/your/dataset/file.jsonl
-```
-
-This must be publically accessible.
-
-## Next steps
-
-Now that you know how to load datasets, you can learn more on how to load your specific dataset format into your target output format [dataset formats docs](dataset-formats).
--- a/docs/dataset_preprocessing.qmd
+++ b/docs/dataset_preprocessing.qmd
@@ -3,11 +3,8 @@ title: Dataset Preprocessing
 description: How datasets are processed
 ---

-## Overview
-
 Dataset pre-processing is the step where Axolotl takes each dataset you've configured alongside
-the [dataset format](dataset-formats) and prompt strategies to:
-
+the (dataset format)[../dataset-formats/] and prompt strategies to:
 - parse the dataset based on the *dataset format*
 - transform the dataset to how you would interact with the model based on the *prompt strategy*
 - tokenize the dataset based on the configured model & tokenizer
@@ -15,12 +12,10 @@ the [dataset format](dataset-formats) and prompt strategies to:

 The processing of the datasets can happen one of two ways:

-1. Before kicking off training by calling `axolotl preprocess config.yaml --debug`
+1. Before kicking off training by calling `python -m axolotl.cli.preprocess /path/to/your.yaml --debug`
 2. When training is started

-### What are the benefits of pre-processing?
-
-When training interactively or for sweeps
+What are the benefits of pre-processing? When training interactively or for sweeps
 (e.g. you are restarting the trainer often), processing the datasets can oftentimes be frustratingly
 slow. Pre-processing will cache the tokenized/formatted datasets according to a hash of dependent
 training parameters so that it will intelligently pull from its cache when possible.
@@ -33,12 +28,8 @@ default path of `./last_run_prepared/`, but will ignore anything already cached
 setting `dataset_prepared_path: ./last_run_prepared`, the trainer will use whatever pre-processed
 data is in the cache.

-### What are the edge cases?
-
-Let's say you are writing a custom prompt strategy or using a user-defined
+What are the edge cases? Let's say you are writing a custom prompt strategy or using a user-defined
 prompt template. Because the trainer cannot readily detect these changes, we cannot change the
-calculated hash value for the pre-processed dataset.
-
-If you have `dataset_prepared_path: ...` set
+calculated hash value for the pre-processed dataset. If you have `dataset_prepared_path: ...` set
 and change your prompt templating logic, it may not pick up the changes you made and you will be
 training over the old prompt.
--- a/docs/debugging.qmd
+++ b/docs/debugging.qmd
@@ -31,13 +31,11 @@ While debugging it's helpful to simplify your test scenario as much as possible.
    - Set `CUDA_VISIBLE_DEVICES` to a single GPU, ex: `export CUDA_VISIBLE_DEVICES=0`.
    - Set `dataset_processes: 1` in your axolotl config or run the training command with `--dataset_processes=1`.
 2. **Use a small dataset**: Construct or use a small dataset from HF Hub. When using a small dataset, you will often have to make sure `sample_packing: False` and `eval_sample_packing: False` to avoid errors.  If you are in a pinch and don't 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):
-
    ```yaml
-    datasets:
+    dataset:
        ...
        shards: 20
    ```
-
 3. **Use a small model**: A good example of a small model is [TinyLlama/TinyLlama-1.1B-Chat-v1.0](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0).
 4. **Minimize iteration time**: Make sure the training loop finishes as fast as possible, with these settings.
    - `micro_batch_size: 1`
@@ -87,7 +85,7 @@ The easiest way to get started is to modify the [.vscode/launch.json](../.vscode

 For example, to mimic the command `cd devtools && CUDA_VISIBLE_DEVICES=0 accelerate launch -m axolotl.cli.train dev_chat_template.yml`, you would use the below configuration[^1].  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 `devtools` and set the `env` variable `HF_HOME` 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.

-```json
+```jsonc
 // .vscode/launch.json
 {
    "version": "0.2.0",
@@ -134,7 +132,7 @@ For example, to mimic the command `cd devtools && CUDA_VISIBLE_DEVICES=0 acceler

 Below is the [./vscode/tasks.json](../.vscode/tasks.json) file that defines the `cleanup-for-dataprep` 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 `cleanup-for-dataprep` 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 `preLaunchTask` argument of the `launch.json` file.

-```json
+```jsonc
 // .vscode/tasks.json
 // this file is used by launch.json
 {
--- a/docs/docker.qmd
+++ b/docs/docker.qmd
@@ -1,146 +0,0 @@
---
-title: "Docker"
-format:
-  html:
-    toc: true
-    toc-depth: 4
---
-
-This section describes the different Docker images that are released by AxolotlAI at [Docker Hub](https://hub.docker.com/u/axolotlai).
-
-::: {.callout-important}
-For Blackwell GPUs, please use the tags with PyTorch 2.7.1 and CUDA 12.8.
-:::
-
-## Base
-
-The base image is the most minimal image that can install Axolotl. It is based on the `nvidia/cuda` image. It includes python, torch, git, git-lfs, awscli, pydantic, and more.
-
-#### Image
-
-```
-axolotlai/axolotl-base
-```
-
-Link: [Docker Hub](https://hub.docker.com/r/axolotlai/axolotl-base)
-
-#### Tags format
-
-```bash
-main-base-py{python_version}-cu{cuda_version}-{pytorch_version}
-```
-
-Tags examples:
-
- `main-base-py3.11-cu128-2.7.1`
- `main-base-py3.11-cu126-2.7.1`
- `main-base-py3.11-cu126-2.7.0`
- `main-base-py3.11-cu126-2.6.0`
- `main-base-py3.11-cu124-2.6.0`
-
-## Main
-
-The main image is the image that is used to run Axolotl. It is based on the `axolotlai/axolotl-base` image and includes the Axolotl codebase, dependencies, and more.
-
-#### Image
-
-```
-axolotlai/axolotl
-```
-
-Link: [Docker Hub](https://hub.docker.com/r/axolotlai/axolotl)
-
-#### Tags format {#sec-main-tags}
-
-```bash
-# on push to main
-main-py{python_version}-cu{cuda_version}-{pytorch_version}
-
-# latest main (currently torch 2.6.0, python 3.11, cuda 12.4)
-main-latest
-
-# nightly build
-{branch}-{date_in_YYYYMMDD}-py{python_version}-cu{cuda_version}-{pytorch_version}
-
-# tagged release
-{version}
-```
-
-:::{.callout-tip}
-
-There may be some extra tags appended to the image, like `-vllm` which installs those packages.
-
-:::
-
-Tags examples:
-
- `main-py3.11-cu128-2.7.1`
- `main-py3.11-cu126-2.7.1`
- `main-py3.11-cu126-2.7.0`
- `main-py3.11-cu126-2.6.0`
- `main-py3.11-cu124-2.6.0`
- `main-latest`
- `main-20250303-py3.11-cu124-2.6.0`
- `main-20250303-py3.11-cu126-2.6.0`
- `0.10.1`
-
-## Cloud
-
-The cloud image is the image that is used to run Axolotl in the cloud. It is based on the `axolotlai/axolotl` image and sets ENV variables like HuggingFace cache directories for volume mounts, tmux, and more for different cloud providers.
-
-:::{.callout-tip}
-
-Jupyter lab is run by default. Set `JUPYTER_DISABLE=1` in the environment variables to disable it.
-
-:::
-
-#### Image
-
-```
-axolotlai/axolotl-cloud
-```
-
-Link: [Docker Hub](https://hub.docker.com/r/axolotlai/axolotl-cloud)
-
-#### Tags format
-
-This uses the same tags as the [`main` image](#sec-main-tags).
-
-#### Environment variables
-
- `JUPYTER_DISABLE`: Disable Jupyter lab.
- `JUPYTER_PASSWORD`: Set a password for the Jupyter lab.
- `PUBLIC_KEY` / `SSH_KEY`: Add a public key for the SSH service.
-
-#### Volume mounts
-
-:::{.callout-tip}
-
-We recommend mounting volumes to `/workspace/data` for data persistence. `/workspace/axolotl` contains the source code and is ephemeral.
-
-:::
-
- `/workspace/data/axolotl-artifacts`: Directory to store Axolotl artifacts.
- `/workspace/data/huggingface-cache`: Directory to store HuggingFace cache.
-
-## Cloud-no-tmux
-
-This is the same as the [`cloud` image](#sec-cloud) but without tmux.
-
-#### Image
-
-```
-axolotlai/axolotl-cloud-term
-```
-
-Link: [Docker Hub](https://hub.docker.com/r/axolotlai/axolotl-cloud-term)
-
-:::{.callout-note}
-
-The naming may be a bit confusing as it has `-term` appended to the end.
-
-:::
-
-#### Tags format
-
-This uses the same tags as the [`cloud` image](#sec-cloud-tags).
--- a/docs/faq.qmd
+++ b/docs/faq.qmd
@@ -3,140 +3,19 @@ title: FAQ
 description: Frequently asked questions
 ---

-### General

 **Q: The trainer stopped and hasn't progressed in several minutes.**

 > A: Usually an issue with the GPUs communicating with each other. See the [NCCL doc](nccl.qmd)

-**Q: exitcode: -9**
+**Q: Exitcode -9**

 > A: This usually happens when you run out of system RAM.

-**Q: exitcode: -7 while using deepspeed**
+**Q: Exitcode -7 while using deepspeed**

 > A: Try upgrading deepspeed w: `pip install -U deepspeed`

 **Q: AttributeError: 'DummyOptim' object has no attribute 'step'**

-**Q: ModuleNotFoundError: No module named 'mpi4py' using single GPU with deepspeed**
-
-> A: You may be using deepspeed with single gpu. Please remove the `deepspeed:` section in the yaml file or `--deepspeed` CLI flag.
-
-**Q: The codes is stuck on saving preprocessed datasets.**
-
-> A: This is usually an issue with the GPU. This can be resolved through setting the os environment variable `CUDA_VISIBLE_DEVICES=0`. If you are on runpod, this is usually a pod issue. Starting a new pod should take care of it.
-
-**Q: Received mismatch error on merge adapters / loading adapters between torch.Size of checkpoint and model.**
-
-> A: This is likely due to vocab size mismatch. By default, Axolotl expands the model's embeddings if the tokenizer has more tokens than the model. Please use the `axolotl merge-lora` command to merge the adapters instead of using your own scripts.
-
-> On the other hand, if the model has more tokens than the tokenizer, Axolotl does not shrink the model's embeddings unless `shrink_embeddings: true` is set in the config.
-
-**Q: How to call Axolotl via custom python scripts?**
-
-> A: Since Axolotl is just Python, please see `src/axolotl/cli/main.py` on how each command is called.
-
-**Q: How to know the value to use for `fsdp_transformer_layer_cls_to_wrap`?**
-
-> A: This is the class name of the transformer layer to wrap with FSDP. For example, for `LlamaForCausalLM`, the value is `LlamaDecoderLayer`. To find this for a specific model, check the model's `PreTrainedModel` definition and look for `_no_split_modules` variable in the `modeling_<model_name>.py` file within `transformers` library.
-
-**Q: ValueError: Asking to pad but the tokenizer does not have a padding token. Please select a token to use as pad_token**
-
-> A: This is because the tokenizer does not have a padding token. Please add a padding token to the tokenizer via:
-
-> ```yaml
-> special_tokens:
->   # str. If you're not sure, set to same as `eos_token`.
->   pad_token: "..."
-> ```
-
-**Q: `IterableDataset error` or `KeyError: 'input_ids'` when using `preprocess` CLI**
-
-> A: This is because you may be using `preprocess` CLI with `pretraining_dataset:` or `skip_prepare_dataset: true` respectively. Please use `axolotl train` CLI directly instead as these datasets are prepared on demand.
-
-**Q: vLLM is not working with Axolotl**
-
-> A: We currently recommend torch 2.6.0 for use with `vllm`. Please ensure you use the right version. For Docker, please use the `main-py3.11-cu124-2.6.0` tag.
-
-**Q: FA2 2.8.0 `undefined symbol` runtime error on CUDA 12.4**
-
-> A: There seems to be a wheel issue with FA2 2.8.0 on CUDA 12.4. Try CUDA 12.6 instead or downgrade to FA2 2.7.4. Please refer to the upstream issue: https://github.com/Dao-AILab/flash-attention/issues/1717.
-
-### Chat templates
-
-**Q: `jinja2.exceptions.UndefinedError: 'dict object' has no attribute 'content' / 'role' / ____`**
-
-> A: This means that the property mapping for the stated attribute does not exist when building `chat_template` prompt. For example, if `no attribute 'content'`, please check you have added the correct mapping for `content` under `message_property_mappings`.
-
-**Q: `Empty template generated for turn ___`**
-
-> A: The `content` is empty for that turn.
-
-**Q: `Could not find content start/end boundary for turn __`**
-
-> A: The specific turn's start/end could not be detected. Please ensure you have set the `eos_token` following your `chat_template`. Otherwise, this could be a `chat_template` which doesn't use proper boundaries for each turn (like system). On the rare occurrence, make sure your content is not `[[dummy_message]]`. Please let us know about this.
-
-**Q: `Content end boundary is before start boundary for turn ___`**
-
-> A: This is an edge case which should not occur. Please create an Issue if this happens.
-
-**Q: `Content end boundary is the same as start boundary for turn ___. This is likely an empty turn.`**
-
-> A: This is likely an empty turn.
-
-**Q: The EOS token is incorrectly being masked or not being masked / `EOS token __ not found in chat template`.**
-
-> A: There can be two reasons:
-
-> 1. This is because of the mismatch between `tokenizer.eos_token` and EOS token in template. Please make sure to set `eos_token: ` under `special_tokens: ` to the same EOS token as in template.
-
-> 2. The EOS token is not in the template. Please check if your template is correct. As an example, `phi_35` template does not use its dedicated EOS token `<|endoftext|>` at the end.
-
-**Q: "`chat_template` choice is `tokenizer_default` but tokenizer's `chat_template` is null. Please add a `chat_template` in tokenizer config"**
-
-> A: This is because the tokenizer does not have a chat template. Please add a chat template in the tokenizer config. See [chat_template](dataset-formats/conversation.qmd#chat-template) for more details.
-
-**Q: The EOT token(s) are incorrectly being masked or not being masked / `EOT token __ not found in chat template`.**
-
-> A: There can be two reasons:
-
-> 1. The EOT token is different from the EOS token and was not specified under `eot_tokens: `. Please set `eot_tokens: ` to the same EOT token(s) as in template.
-
-> 2. There is more than one EOT token per turn in the template. Please raise an issue with examples as we recognize this as an edge case.
-
-**Q: `EOT token encoding failed. Please check if the token is valid and can be encoded.`**
-
-> A: There could be some issue with the tokenizer or unicode encoding. Please raise an issue with examples with the EOT token & tokenizer causing the issue.
-
-**Q: `EOT token __ is encoded as multiple tokens.`**
-
-> A: This is because the EOT token is encoded as multiple tokens which can cause unexpected behavior. Please add it under `tokens: ` or (recommended) override unused added_tokens via `added_tokens_overrides: `.
-
-**Q: `Conflict between train_on_eos and train_on_eot. eos_token is in eot_tokens and train_on_eos != train_on_eot`**
-
-> A: This is because the EOS token is in the `eot_tokens: ` while mismatch between `train_on_eos: ` and `train_on_eot: `. This will cause one to override the other. Please ensure that `train_on_eos: ` and `train_on_eot: ` are the same or remove the EOS token from `eot_tokens: `.
-
-**Q: If `eot_tokens: ` is not provided, what happens?**
-
-> A: If `eot_tokens: ` is not provided, the default behavior is the same as before. EOS tokens used to delimit turns are masked/unmasked depending on whether the turn is trainable.
-
-> Internally, `eot_tokens: tokenizer.eos_token` and `train_on_eot: train_on_eos` (which defaults to `turn`). This transition helps clarify the naming and behavior of EOT/EOS tokens.
-
-**Q: `Data processing error: CAS service error`**
-
-> A: Try disabling XET with `export HF_HUB_DISABLE_XET=1`
-
-**Q: `torch._inductor.exc.LoweringException: NoValidChoicesError: No choices to select, please consider adding ATEN into max_autotune_gemm_backends config (defined in torch/_inductor/config.py) to allow at least one choice. `**
-
-> A: Depending on the version of torch, you may need to include this in your YAML:
-
-> ```yaml
-> flex_attn_compile_kwargs:
->   dynamic: false
->   mode: max-autotune-no-cudagraphs
-> ```
-
-**Q: `ValueError("Backward pass should have cleared tracker of all tensors")`
-
-> A: This may happen due to edge cases in using the modern OffloadActivations context manager for CUDA streams. If you encounter this error, you may have success using the naive implementation with `offload_activations: legacy` in your YAML.
+> A: You may be using deepspeed with single gpu. Please don't set `deepspeed:` in yaml or cli.
--- a/docs/fsdp_qlora.qmd
+++ b/docs/fsdp_qlora.qmd
@@ -20,7 +20,7 @@ To enable `QLoRA` with `FSDP`, you need to perform the following steps:
 > See the [example config](#example-config) file in addition to reading these instructions.

 1. Set `adapter: qlora` in your axolotl config file.
-2. Enable FSDP in your axolotl config, as [described here](multi-gpu.qmd#sec-fsdp).
+2. Enable FSDP in your axolotl config, as [described here](https://github.com/axolotl-ai-cloud/axolotl?tab=readme-ov-file#fsdp).
 3. Use one of the supported model types: `llama`, `mistral` or `mixtral`.

 ## Example Config
--- a/docs/getting-started.qmd
+++ b/docs/getting-started.qmd
@@ -1,186 +0,0 @@
---
-title: "Quickstart"
-format:
-  html:
-    toc: true
-    toc-depth: 3
-    number-sections: true
-execute:
-  enabled: false
---
-
-This guide will walk you through your first model fine-tuning project with Axolotl.
-
-## Quick Example {#sec-quick-example}
-
-Let's start by fine-tuning a small language model using LoRA. This example uses a 1B parameter model to ensure it runs on most GPUs.
-Assuming `axolotl` is installed (if not, see our [Installation Guide](installation.qmd))
-
-1. Download example configs:
-```bash
-axolotl fetch examples
-```
-
-2. Run the training:
-```bash
-axolotl train examples/llama-3/lora-1b.yml
-```
-
-That's it! Let's understand what just happened.
-
-## Understanding the Process {#sec-understanding}
-
-### The Configuration File {#sec-config}
-
-The YAML configuration file controls everything about your training. Here's what (part of) our example config looks like:
-
-```yaml
-base_model: NousResearch/Llama-3.2-1B
-
-load_in_8bit: true
-adapter: lora
-
-datasets:
-  - path: teknium/GPT4-LLM-Cleaned
-    type: alpaca
-dataset_prepared_path: last_run_prepared
-val_set_size: 0.1
-output_dir: ./outputs/lora-out
-```
-
-::: {.callout-tip}
-`load_in_8bit: true` and `adapter: lora` enables LoRA adapter finetuning.
-
- To perform Full finetuning, remove these two lines.
- To perform QLoRA finetuning, replace with `load_in_4bit: true` and `adapter: qlora`.
-:::
-
-See our [config options](config-reference.qmd) for more details.
-
-### Training {#sec-training}
-
-When you run `axolotl train`, Axolotl:
-
-1. Downloads the base model
-2. (If specified) applies QLoRA/LoRA adapter layers
-3. Loads and processes the dataset
-4. Runs the training loop
-5. Saves the trained model and / or LoRA weights
-
-## Your First Custom Training {#sec-custom}
-
-Let's modify the example for your own data:
-
-1. Create a new config file `my_training.yml`:
-
-```yaml
-base_model: NousResearch/Nous-Hermes-llama-1b-v1
-
-load_in_8bit: true
-adapter: lora
-
-# Training settings
-micro_batch_size: 2
-num_epochs: 3
-learning_rate: 0.0003
-
-# Your dataset
-datasets:
-  - path: my_data.jsonl        # Your local data file
-    type: alpaca               # Or other format
-```
-
-This specific config is for LoRA fine-tuning a model with instruction tuning data using
-the `alpaca` dataset format, which has the following format:
-
-```json
-{
-    "instruction": "Write a description of alpacas.",
-    "input": "",
-    "output": "Alpacas are domesticated South American camelids..."
-}
-```
-
-Please see our [Dataset Formats](dataset-formats) for more dataset formats and how to
-format them.
-
-2. Prepare your JSONL data in the specified format (in this case, the expected `alpaca`
-format):
-
-```json
-{"instruction": "Classify this text", "input": "I love this!", "output": "positive"}
-{"instruction": "Classify this text", "input": "Not good at all", "output": "negative"}
-```
-
-3. Run the training:
-
-```bash
-axolotl train my_training.yml
-```
-
-## Common Tasks {#sec-common-tasks}
-
-::: {.callout-tip}
-
-The same yaml file is used for training, inference, and merging.
-
-:::
-
-### Testing Your Model {#sec-testing}
-
-After training, test your model:
-
-```bash
-axolotl inference my_training.yml --lora-model-dir="./outputs/lora-out"
-```
-
-More details can be found in [Inference](inference.qmd).
-
-### Using a UI {#sec-ui}
-
-Launch a Gradio interface:
-
-```bash
-axolotl inference my_training.yml --lora-model-dir="./outputs/lora-out" --gradio
-```
-
-### Preprocessing Data {#sec-preprocessing}
-
-For large datasets, preprocess first:
-
-```bash
-axolotl preprocess my_training.yml
-```
-
-Please make sure to set `dataset_prepared_path: ` in your config to set the path to save the prepared dataset.
-
-More details can be found in [Dataset Preprocessing](dataset_preprocessing.qmd).
-
-### Merging LoRA weights {#sec-merging-lora}
-
-To merge the LoRA weights back into the base model, run:
-
-```bash
-axolotl merge-lora my_training.yml --lora-model-dir="./outputs/lora-out"
-```
-
-The merged model will be saved in the `{output_dir}/merged` directory.
-
-More details can be found in [Merging LoRA weights](inference.qmd#sec-merging).
-
-## Next Steps {#sec-next-steps}
-
-Now that you have the basics, you might want to:
-
- Try different model architectures
- Experiment with hyperparameters
- Use more advanced training methods
- Scale up to larger models
-
-Check our other guides for details on these topics:
-
- [Configuration Guide](config-reference.qmd) - Full configuration options
- [Dataset Loading](dataset_loading.qmd) - Loading datasets from various sources
- [Dataset Formats](dataset-formats) - Working with different data formats
- [Multi-GPU Training](multi-gpu.qmd)
- [Multi-Node Training](multi-node.qmd)
--- a/docs/gradient_checkpointing.qmd
+++ b/docs/gradient_checkpointing.qmd
@@ -1,29 +0,0 @@
---
-title: Gradient Checkpointing and Activation Offloading
---
-
-Gradient checkpointing and activation offloading are techniques used to optimize the performance of deep learning
-models by reducing the memory footprint and improving computational efficiency.
-
-### Enabling Gradient Checkpointing
-
-```yaml
-gradient_checkpointing: true
-```
-
-### Enabling Activation Offloading
-
-```yaml
-gradient_checkpointing: true  # required for activation offloading
-activation_offloading: true
-```
-
-Activation offloading variants:
-
-The default `activation_offloading: true` offloads activations to CPU and uses CUDA streams
-to overlap the communications and computations when offloading.
-
-The `activation_offloading: legacy` naively offloads activations to CPU and without additional optimizations.
-
-For resource constrained environments with limited CPU memory, `activation_offloading: disk` offloads
-activations to disk instead of CPU RAM so that much larger context lengths can be trained with minimal memory.
--- a/docs/images/ray-cluster-dashboard.png
+++ b/docs/images/ray-cluster-dashboard.png
--- a/docs/inference.qmd
+++ b/docs/inference.qmd
@@ -1,151 +0,0 @@
---
-title: "Inference and Merging"
-format:
-  html:
-    toc: true
-    toc-depth: 3
-    number-sections: true
-execute:
-  enabled: false
---
-
-This guide covers how to use your trained models for inference, including model loading, interactive testing, merging adapters, and common troubleshooting steps.
-
-## Quick Start {#sec-quickstart}
-
-::: {.callout-tip}
-Use the same config used for training on inference/merging.
-:::
-
-### Basic Inference {#sec-basic}
-
-::: {.panel-tabset}
-
-## LoRA Models
-
-```{.bash}
-axolotl inference your_config.yml --lora-model-dir="./lora-output-dir"
-```
-
-## Full Fine-tuned Models
-
-```{.bash}
-axolotl inference your_config.yml --base-model="./completed-model"
-```
-
-:::
-
-## Advanced Usage {#sec-advanced}
-
-### Gradio Interface {#sec-gradio}
-
-Launch an interactive web interface:
-
-```{.bash}
-axolotl inference your_config.yml --gradio
-```
-
-### File-based Prompts {#sec-file-prompts}
-
-Process prompts from a text file:
-
-```{.bash}
-cat /tmp/prompt.txt | axolotl inference your_config.yml \
-  --base-model="./completed-model" --prompter=None
-```
-
-### Memory Optimization {#sec-memory}
-
-For large models or limited memory:
-
-```{.bash}
-axolotl inference your_config.yml --load-in-8bit=True
-```
-
-## Merging LoRA Weights {#sec-merging}
-
-Merge LoRA adapters with the base model:
-
-```{.bash}
-axolotl merge-lora your_config.yml --lora-model-dir="./completed-model"
-```
-
-### Memory Management for Merging {#sec-memory-management}
-
-::: {.panel-tabset}
-
-## Configuration Options
-
-```{.yaml}
-gpu_memory_limit: 20GiB  # Adjust based on your GPU
-lora_on_cpu: true        # Process on CPU if needed
-```
-
-## Force CPU Merging
-
-```{.bash}
-CUDA_VISIBLE_DEVICES="" axolotl merge-lora ...
-```
-
-:::
-
-## Tokenization {#sec-tokenization}
-
-### Common Issues {#sec-tokenization-issues}
-
-::: {.callout-warning}
-Tokenization mismatches between training and inference are a common source of problems.
-:::
-
-To debug:
-
-1. Check training tokenization:
-```{.bash}
-axolotl preprocess your_config.yml --debug
-```
-
-2. Verify inference tokenization by decoding tokens before model input
-
-3. Compare token IDs between training and inference
-
-### Special Tokens {#sec-special-tokens}
-
-Configure special tokens in your YAML:
-
-```{.yaml}
-special_tokens:
-  bos_token: "<s>"
-  eos_token: "</s>"
-  unk_token: "<unk>"
-tokens:
-  - "<|im_start|>"
-  - "<|im_end|>"
-```
-
-## Troubleshooting {#sec-troubleshooting}
-
-### Common Problems {#sec-common-problems}
-
-::: {.panel-tabset}
-
-## Memory Issues
-
- Use 8-bit loading
- Reduce batch sizes
- Try CPU offloading
-
-## Token Issues
-
- Verify special tokens
- Check tokenizer settings
- Compare training and inference preprocessing
-
-## Performance Issues
-
- Verify model loading
- Check prompt formatting
- Ensure temperature/sampling settings
-
-:::
-
-For more details, see our [debugging guide](debugging.qmd).
--- a/docs/input_output.qmd
+++ b/docs/input_output.qmd
@@ -3,4 +3,263 @@ title: Template-free prompt construction
 description: "Template-free prompt construction with the `input_output` format"
 ---

-The documentation moved to [here](dataset-formats/template_free.qmd).
+<!-- TOC -->
+
+- [Background](#background)
+    - [Masking Inputs](#masking-inputs)
+    - [You may not want prompt templates](#you-may-not-want-prompt-templates)
+    - [The `input_output` format](#the-input_output-format)
+- [Usage](#usage)
+    - [1. Prepare Data](#1-prepare-data)
+    - [2. Use `type: input_output`](#2-use-type-input_output)
+    - [3. Check the prompts](#3-check-the-prompts)
+
+<!-- /TOC -->
+
+<a id="markdown-background" name="background"></a>
+
+## Background
+
+<a id="markdown-masking-inputs" name="masking-inputs"></a>
+
+### Masking Inputs
+
+One of the most popular features of
+[axolotl](https://github.com/axolotl-ai-cloud/axolotl) is
+setting the following configuration value:
+
+
+```yaml
+train_on_inputs: false
+```
+
+If you declare a [dataset formats](https://github.com/axolotl-ai-cloud/axolotl?tab=readme-ov-file#dataset)
+such as `alpaca` or `chatml`, axolotl knows what is an input
+(i.e. human) vs. an output (i.e. the assistant) and masks the input
+labels so that your model can focus on predicting the outputs only.
+
+<a id="markdown-you-may-not-want-prompt-templates" name="you-may-not-want-prompt-templates"></a>
+
+### You may not want prompt templates
+
+However, there are many situations where you don't want to use one of
+these formats or templates. This is because they can:
+
+-   Add unnecessary boilerplate to your prompts.
+-   Create artifacts like special delimiters `<|im_start|>` that can
+    quickly become footguns if you don't include them correctly at
+    inference time.
+-   Enforce a *chat* 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.
+-   Limit you to only certain roles that the template allows.
+
+<a id="markdown-the-inputoutput-format" name="the-inputoutput-format"></a>
+
+### The `input_output` format
+
+You can construct your prompts without a template by using the
+`input_output` format, by setting `type: input_output` in your
+configuration file like this:
+
+**config.yml**
+
+```yaml
+train_on_inputs: false # Mask segments of your data
+datasets:
+  - path: output.jsonl
+    type: input_output  # use template free prompt construction
+```
+
+Unlike `type: completion`, which is also template-free,
+`type: input_output` allows you to mask segments of your text. More
+details on how this works are described below.
+
+<a id="markdown-usage" name="usage"></a>
+
+## Usage
+
+This is how you can use the `input_output` format:
+
+<a id="markdown-1-prepare-data" name="1-prepare-data"></a>
+
+### 1. Prepare Data
+
+To use the `input_output` format, collect your data in the following
+format into a jsonl file (below is the first row from the file
+`output`.jsonl` pretty printed):
+
+```bash
+$ head -n1 output.jsonl | python -m json.tool
+```
+
+:::{.cell-output .cell-output-stdout}
+    {
+        "segments": [
+            {
+                "label": true,
+                "text": "<s>Hello\n"
+            },
+            {
+                "label": true,
+                "text": "hi there!. "
+            },
+            {
+                "label": false,
+                "text": "goodbye "
+            },
+            {
+                "label": true,
+                "text": "farewell</s>"
+            }
+        ]
+    }
+:::
+
+Set `label:false` when you want to mask a segment of text so that the
+model isn't trained on it. Some things to keep in mind:
+
+> [!IMPORTANT]
+> 1.  **EOS, BOS, spaces, newlines etc. are entirely up to you. Axolotl
+    concatenates all the segments as-is.** The tokenizer doesn't add
+    anything additional. Notice how I added spaces, newlines, `<s>`
+    (BOS), and `</s>` (EOS) myself.
+> 2.  Make sure you check the materialized output to validate that the
+    prompt is getting assembled how you like.
+
+<a id="markdown-2-use-type-inputoutput" name="2-use-type-inputoutput"></a>
+
+### 2. Use `type: input_output`
+
+Let's materialize data with our `output.jsonl` file by setting
+`type: input_output` in our axolotl config:
+
+```yaml
+# training_config.yaml
+base_model: mistralai/Mistral-7B-v0.1
+data_seed: 49
+seed: 49
+
+datasets:
+  - path: output.jsonl
+    type: input_output
+val_set_size: 0.1
+
+sequence_len: 896
+sample_packing: false
+
+micro_batch_size: 2
+gradient_accumulation_steps: 3
+eval_batch_size: 2
+num_epochs: 1
+learning_rate: 0.0002
+
+train_on_inputs: false
+special_tokens:
+  bos_token: "<s>"
+  eos_token: "</s>"
+  unk_token: "<unk>"
+```
+
+You can use the following command to materialize your data. The
+`--debug` flag will print the tokens, along with the labels so you can
+verify that the correct items are being ignored:
+
+```bash
+$ python -m axolotl.cli.preprocess training_config.yaml --debug
+
+...
+[2024-03-05 23:36:46,969] [INFO] [axolotl.check_example_labels:35] [PID:607731] [RANK:0] <s>(1, 1) Hello(22557, 22557)
+(13, 13) hi(12014, 12014) there(736, 736) !(28808, 28808) .(28723, 28723) (28705, 28705) good(-100, 1179) bye(-100, 17664) (-100, 28705) fare(19111, 19111) well(5458, 5458) </s>(2, 2)
+
+```
+
+The format is `decoded_token`(`label`, `token_id`), for example,
+`<s>(1, 1)` means that the token is `<s>`, the label is `1` and the
+token_id is `1`. When the label is `-100` then that token is ignored for
+training.
+
+<a id="markdown-3-check-the-prompts" name="3-check-the-prompts"></a>
+
+### 3. Check the prompts
+
+Here is another way to check the materialized output:
+
+```python
+from transformers import AutoTokenizer
+from datasets import load_from_disk
+import yaml
+
+directory = !ls last_run_prepared/
+with open('training_config.yaml', 'r') as f:
+    cfg = yaml.safe_load(f)
+model_id = cfg['base_model']
+tok = AutoTokenizer.from_pretrained(model_id)
+ds = load_from_disk(f'last_run_prepared/{directory[0]}/')
+```
+
+```python
+>>> row = ds[0]
+>>> print(tok.decode(row['input_ids']))
+<s> Hello
+    hi there!.  goodbye  farewell</s>
+```
+
+We can check that the right tokens are ignored by comparing the labels
+to each token:
+
+```python
+import pandas as pd
+pd.DataFrame([{'token': tok.decode(i), 'label': l, 'id':i} for i,l in
+              zip(row['input_ids'], row['labels'])])
+```
+
+| token | label | id    |
+|-------|-------|-------|
+| 0     | \<s\> | 1     |
+| 1     | Hello | 22557 |
+| 2     | \\n   | 13    |
+| 3     | hi    | 12014 |
+| 4     | there | 736   |
+| 5     | !     | 28808 |
+| 6     | .     | 28723 |
+| 7     |       | 28705 |
+| 8     | good  | -100  |
+| 9     | bye   | -100  |
+| 10    |       | -100  |
+| 11    | fare  | 19111 |
+| 12    | well  | 5458  |
+| 13    | \</s\>| 2     |
+
+
+
+If we look at the input data, the above table seems correct! (The jsonl
+version is repeated below for reference):
+
+
+```bash
+$ head -n1 output.jsonl | python -m json.tool
+```
+
+:::{.cell-output .cell-output-stdout}
+    {
+        "segments": [
+            {
+                "label": true,
+                "text": "<s>Hello\n"
+            },
+            {
+                "label": true,
+                "text": "hi there!. "
+            },
+            {
+                "label": false,
+                "text": "goodbye "
+            },
+            {
+                "label": true,
+                "text": "farewell</s>"
+            }
+        ]
+    }
+:::
--- a/docs/installation.qmd
+++ b/docs/installation.qmd
@@ -1,173 +0,0 @@
---
-title: "Installation"
-format:
-  html:
-    toc: true
-    toc-depth: 3
-    number-sections: true
-execute:
-  enabled: false
---
-
-This guide covers all the ways you can install and set up Axolotl for your environment.
-
-## Requirements {#sec-requirements}
-
- NVIDIA GPU (Ampere architecture or newer for `bf16` and Flash Attention) or AMD GPU
- Python ≥3.11
- PyTorch ≥2.6.0
-
-## Installation Methods {#sec-installation-methods}
-
-::: {.callout-important}
-Please make sure to have Pytorch installed before installing Axolotl in your local environment.
-
-Follow the instructions at: [https://pytorch.org/get-started/locally/](https://pytorch.org/get-started/locally/)
-:::
-
-::: {.callout-important}
-For Blackwell GPUs, please use Pytorch 2.7.0 and CUDA 12.8.
-:::
-
-### PyPI Installation (Recommended) {#sec-pypi}
-
-```{.bash}
-pip3 install -U packaging setuptools wheel ninja
-pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]
-```
-
-We use `--no-build-isolation` in order to detect the installed PyTorch version (if
-installed) in order not to clobber it, and so that we set the correct version of
-dependencies that are specific to the PyTorch version or other installed
-co-dependencies.
-
-### uv Installation {#sec-uv}
-
-uv is a fast, reliable Python package installer and resolver built in Rust. It offers significant performance improvements over pip and provides better dependency resolution, making it an excellent choice for complex environments.
-
-Install uv if not already installed
-```{.bash}
-curl -LsSf https://astral.sh/uv/install.sh | sh
-source $HOME/.local/bin/env
-```
-
-Choose your CUDA version to use with PyTorch; e.g. `cu124`, `cu126`, `cu128`,
-then create the venv and activate
-```{.bash}
-export UV_TORCH_BACKEND=cu126
-uv venv --no-project --relocatable
-source .venv/bin/activate
-```
-
-Install PyTorch
- PyTorch 2.6.0 recommended
-```{.bash}
-uv pip install packaging setuptools wheel
-uv pip install torch==2.6.0
-uv pip install awscli pydantic
-```
-
-Install axolotl from PyPi
-```{.bash}
-uv pip install --no-build-isolation axolotl[deepspeed,flash-attn]
-
-# optionally install with vLLM if you're using torch==2.6.0 and want to train w/ GRPO
-uv pip install --no-build-isolation axolotl[deepspeed,flash-attn,vllm]
-```
-
-### Edge/Development Build {#sec-edge-build}
-
-For the latest features between releases:
-
-```{.bash}
-git clone https://github.com/axolotl-ai-cloud/axolotl.git
-cd axolotl
-pip3 install -U packaging setuptools wheel ninja
-pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'
-```
-
-### Docker {#sec-docker}
-
-```{.bash}
-docker run --gpus '"all"' --rm -it axolotlai/axolotl:main-latest
-```
-
-For development with Docker:
-
-```{.bash}
-docker compose up -d
-```
-
-::: {.callout-tip}
-### Advanced Docker Configuration
-```{.bash}
-docker run --privileged --gpus '"all"' --shm-size 10g --rm -it \
-  --name axolotl --ipc=host \
-  --ulimit memlock=-1 --ulimit stack=67108864 \
-  --mount type=bind,src="${PWD}",target=/workspace/axolotl \
-  -v ${HOME}/.cache/huggingface:/root/.cache/huggingface \
-  axolotlai/axolotl:main-latest
-```
-:::
-
-::: {.callout-important}
-For Blackwell GPUs, please use `axolotlai/axolotl:main-py3.11-cu128-2.7.0` or the cloud variant `axolotlai/axolotl-cloud:main-py3.11-cu128-2.7.0`.
-:::
-
-Please refer to the [Docker documentation](docker.qmd) for more information on the different Docker images that are available.
-
-## Cloud Environments {#sec-cloud}
-
-### Cloud GPU Providers {#sec-cloud-gpu}
-
-For providers supporting Docker:
-
- Use `axolotlai/axolotl-cloud:main-latest`
- Available on:
-    - [RunPod](https://runpod.io/gsc?template=v2ickqhz9s&ref=6i7fkpdz)
-    - [Vast.ai](https://cloud.vast.ai?ref_id=62897&template_id=bdd4a49fa8bce926defc99471864cace&utm_source=axolotl&utm_medium=partner&utm_campaign=template_launch_july2025&utm_content=docs_link)
-    - [PRIME Intellect](https://app.primeintellect.ai/dashboard/create-cluster?image=axolotl&location=Cheapest&security=Cheapest&show_spot=true)
-    - [Modal](https://www.modal.com?utm_source=github&utm_medium=github&utm_campaign=axolotl)
-    - [Novita](https://novita.ai/gpus-console?templateId=311)
-    - [JarvisLabs.ai](https://jarvislabs.ai/templates/axolotl)
-    - [Latitude.sh](https://latitude.sh/blueprint/989e0e79-3bf6-41ea-a46b-1f246e309d5c)
-
-### Google Colab {#sec-colab}
-
-Use our [example notebook](../examples/colab-notebooks/colab-axolotl-example.ipynb).
-
-## Platform-Specific Instructions {#sec-platform-specific}
-
-### macOS {#sec-macos}
-
-```{.bash}
-pip3 install --no-build-isolation -e '.'
-```
-
-See @sec-troubleshooting for Mac-specific issues.
-
-### Windows {#sec-windows}
-
-::: {.callout-important}
-We recommend using WSL2 (Windows Subsystem for Linux) or Docker.
-:::
-
-## Environment Managers {#sec-env-managers}
-
-### Conda/Pip venv {#sec-conda}
-
-1. Install Python ≥3.11
-2. Install PyTorch: https://pytorch.org/get-started/locally/
-3. Install Axolotl:
-   ```{.bash}
-   pip3 install -U packaging setuptools wheel ninja
-   pip3 install --no-build-isolation -e '.[flash-attn,deepspeed]'
-   ```
-4. (Optional) Login to Hugging Face:
-   ```{.bash}
-   huggingface-cli login
-   ```
-
-## Troubleshooting {#sec-troubleshooting}
-
-If you encounter installation issues, see our [FAQ](faq.qmd) and [Debugging Guide](debugging.qmd).
--- a/docs/lora_optims.qmd
+++ b/docs/lora_optims.qmd
@@ -1,136 +0,0 @@
---
-title: "LoRA Optimizations"
-description: "Custom autograd functions and Triton kernels in Axolotl for optimized LoRA fine-tuning"
---
-
-Inspired by [Unsloth](https://github.com/unslothai/unsloth), we've implemented two
-optimizations for LoRA and QLoRA fine-tuning, supporting both single GPU and multi-GPU
-(in the DDP and DeepSpeed settings) training. These include (1) SwiGLU and GEGLU activation function
-Triton kernels, and (2) LoRA MLP and attention custom autograd functions. Our goal was
-to leverage operator fusion and tensor re-use in order to improve speed and reduce
-memory usage during the forward and backward passes of these calculations.
-
-We currently support several common model architectures, including (but not limited to):
-
- `llama`
- `mistral`
- `qwen2`
- `gemma`
- `gemma2`
- `gemma3`
-
-<details>
-
-The set of models we support is currently limited by our attention patching strategy,
-which assumes (and replaces) specific code blocks for query / key / value and output
-projections:
-
-```python
-ORIGINAL_QKV_CODE = """
-    query_states = self.q_proj(hidden_states).view(hidden_shape).transpose(1, 2)
-    key_states = self.k_proj(hidden_states).view(hidden_shape).transpose(1, 2)
-    value_states = self.v_proj(hidden_states).view(hidden_shape).transpose(1, 2)
-""".lstrip(
-    "\n"
-)
-
-ORIGINAL_O_CODE = """
-    attn_output = self.o_proj(attn_output)
-""".lstrip(
-    "\n"
-)
-```
-
-Is replaced with:
-
-```python
-PATCHED_QKV_CODE = """
-    query_states, key_states, value_states = self.apply_qkv(hidden_states)
-    query_states = query_states.view(hidden_shape).transpose(1, 2)
-    key_states = key_states.view(hidden_shape).transpose(1, 2)
-    value_states = value_states.view(hidden_shape).transpose(1, 2)
-""".lstrip(
-    "\n"
-)
-
-PATCHED_O_CODE = """
-    attn_output = self.apply_o(attn_output)
-""".lstrip(
-    "\n"
-)
-```
-
-Where `apply_qkv` and `apply_o` are defined in the `axolotl.kernels.lora` module.
-
-We welcome testing of other model architectures and / or PRs to expand our patching
-logic to be compatible with more of them.
-
-</details>
-
-::: {.callout-tip}
-Check out our [LoRA optimizations blog](https://axolotlai.substack.com/p/accelerating-lora-fine-tuning-with).
-:::
-
-## Usage
-
-These optimizations can be enabled in your Axolotl config YAML file. The
-`lora_mlp_kernel` option enables the optimized MLP path, while `lora_qkv_kernel` and
-`lora_o_kernel` enable the fused query-key-value projection and optimized output
-projection, respectively.
-
-```yaml
-lora_mlp_kernel: true
-lora_qkv_kernel: true
-lora_o_kernel: true
-```
-
-::: {.callout-note}
-Currently, LoRA kernels are not supported for RLHF training, only SFT.
-:::
-
-## Requirements
-
- One or more NVIDIA or AMD GPUs (in order to use the Triton kernels)
-    - Note: Set `TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL=1` to enable [memory-efficient attention on AMD GPUs](https://github.com/ROCm/aotriton/issues/16#issuecomment-2346675491)
- Targeted LoRA adapters cannot use Dropout
-    - This may limit model expressivity / cause overfitting
- Targeted LoRA adapters cannot have bias terms
-    - This may limit model expressivity
-
-Models with pre-existing LoRA adapters that use Dropout or have bias terms may need to
-be re-finetuned without these features in order to be useful.
-
-## Implementation details
-
-### Custom autograd functions
-
-The LoRA MLP autograd function optimizes the entire MLP computation path. It fuses the
-LoRA and base weight computations together and provides a single, efficient backward
-pass for the entire MLP block.
-
-For attention components, similar optimizations are provided through a function that
-handles the query, key, and value projections, and a function that handles the output
-projection. They are designed to work with the existing `transformers` attention
-implementation via some monkey-patching logic.
-
-### Triton kernels
-
-Two activation functions (SwiGLU and GeGLU) are implemented with Triton kernels for
-improved speed and memory performance. These kernels handle both the forward and
-backward passes.
-
-### Integration
-
-The custom autograd functions and Triton kernels are designed to work together. The
-autograd function manages the high-level computation flow and gradient tracking, while
-calling the Triton kernels for the activation function computation. During the backward
-pass, the kernel computes both the activation output and the required gradients, which
-the autograd function then uses to compute the final gradients for the entire
-computation path.
-
-## Future Work
-
- Support for additional model architectures
- Support for the FSDP setting
- Support for dropout and bias
- Additional operator fusions
--- a/docs/lr_groups.qmd
+++ b/docs/lr_groups.qmd
@@ -1,29 +0,0 @@
---
-title: Learning Rate Groups
-description: "Setting different learning rates by module name"
---
-
-## Background
-
-Inspired by LoRA+, Axolotl allows practitioners to specify separate learning rates for each module or groups of
-modules in a model.
-
-## Example
-
-```yaml
-lr_groups:
-  - name: o_proj
-    modules:
-      - self_attn.o_proj.weight
-    lr: 1e-6
-  - name: q_proj
-    modules:
-      - model.layers.2.self_attn.q_proj.weight
-    lr: 1e-5
-
-learning_rate: 2e-5
-```
-
-In this example, we have a default learning rate of 2e-5 across the entire model, but we have a separate learning rate
-of 1e-6 for all the self attention `o_proj` modules across all layers, and a learning are of 1e-5 to the 3rd layer's
-self attention `q_proj` module.
--- a/docs/mac.qmd
+++ b/docs/mac.qmd
@@ -19,5 +19,4 @@ Current support:
 - [ ] DeepSpeed

 Untested:
-
 - FSDP
--- a/docs/mixed_precision.qmd
+++ b/docs/mixed_precision.qmd
@@ -1,149 +0,0 @@
---
-title: "Mixed Precision Training"
-format:
-  html:
-    toc: true
-    toc-depth: 3
-    number-sections: true
-    code-tools: true
-execute:
-  enabled: false
---
-
-Mixed precision training uses lower precision data types to reduce memory usage and increase training speed while maintaining model quality. Axolotl supports several mixed precision formats:
-
- **FP16** - Half precision 16-bit (Pascal generation+)
- **BF16** - Brain Float 16-bit (Ampere generation+)
- **FP8** - 8-bit floating point (Hopper generation+)
-
-## FP16 Mixed Precision {#sec-fp16}
-
-### Overview {#sec-fp16-overview}
-
-FP16 is the traditional half-precision format, supported on older GPUs but can be less numerically stable than BF16.
-
-### Configuration {#sec-fp16-config}
-
-```{.yaml}
-fp16: true
-```
-
-### FP16 Considerations {#sec-fp16-considerations}
-
- May require gradient scaling to prevent underflow
- Less numerically stable than BF16
- Can cause training instability with some model architectures
- Consider using BF16 if your hardware supports it
-
-## BF16 Mixed Precision {#sec-bf16}
-
-### Overview {#sec-bf16-overview}
-
-BF16 (Brain Float 16) offers better numerical stability than FP16 and is the recommended mixed precision format for modern GPUs. It provides the same dynamic range as FP32 while using half the memory.
-
-### Configuration {#sec-bf16-config}
-
-```{.yaml}
-# Automatic BF16 detection (recommended)
-bf16: auto
-
-# Or explicitly enable
-bf16: true
-
-# For evaluation with BF16
-bf16: full  # Equivalent to bf16_full_eval in the HF trainer
-```
-
-## FP8 Mixed Precision {#sec-fp8}
-
-::: {.callout-note}
-FP8 support is experimental and requires compatible hardware (H100, H200) and recent PyTorch versions with TorchAO.
-:::
-
-### What is FP8? {#sec-fp8-overview}
-
-FP8 (8-bit floating point) can provide significant time savings compared to FP16/BF16 while maintaining training stability. Axolotl's implementation uses PyTorch's TorchAO library with "tensorwise" scaling strategy.
-
-### Requirements {#sec-fp8-software}
-
- Hopper+ GPUs (H100/H200)
- PyTorch 2.7+ (+ compatible TorchAO version)
- CUDA 12.4+
-
-### Configuration {#sec-fp8-config}
-
-Add to your YAML config:
-
-```{.yaml}
-# Enable FP8 mixed precision
-fp8: true
-
-# Optional: Enable FP8 for FSDP all-gather operations
-fp8_enable_fsdp_float8_all_gather: true
-
-# Enable torch.compile (almost always necessary for FP8 speedups)
-torch_compile: true
-```
-
-::: {.callout-important}
-**torch.compile is critical for FP8 performance**
-
-FP8 training requires `torch_compile: true` to see meaningful speedups. Without compilation, FP8 may actually be slower and use more memory than FP16/BF16.
-:::
-
-### Advanced FP8 Configs {#sec-fp8-advanced}
-
-For [FSDP](multi-gpu.qmd#sec-fsdp) (Fully Sharded Data Parallel) training:
-
-```{.yaml}
-fp8: true
-fp8_enable_fsdp_float8_all_gather: true
-
-torch_compile: true
-
-# FSDP configuration
-fsdp_version: 2
-fsdp_config:
-  offload_params: false
-  cpu_ram_efficient_loading: true
-  auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  transformer_layer_cls_to_wrap: LlamaDecoderLayer
-  state_dict_type: FULL_STATE_DICT
-  reshard_after_forward: true
-```
-
-## Best Practices {#sec-best-practices}
-
-### Choosing Precision Format {#sec-choosing-format}
-
- **Start with automatic detection**: `bf16: auto`
- **For Hopper+ (H100/H200)**: Try FP8 + torch.compile for maximum speed
- **For Ampere (A100/RTX 30/40)**: Use BF16
- **For older Pascal/Turing GPUs**: Use FP16 with caution
- **For very old or unsupported GPUs**: Use FP32
-
-### Validation and Testing {#sec-validation}
-
-Always validate your mixed precision setup:
-
- **Start with a small dataset** to verify stability
- **Monitor loss curves** for irregularities
- **Compare with FP32 baseline** when possible
- **Test evaluation metrics** match expectations
-
-### FP8 Particulars {#sec-fp8-details}
-
- Use cases
-  - Single GPU training
-  - Multi GPU training with FSDP2 or Deepspeed
- Speedups
-  - Please refer to the [TorchAO FP8 training benchmarks](https://github.com/pytorch/ao/tree/main/torchao/float8#rowwise-scaling) for expected matmul speedups for different (M, K, N) settings
-  - Concrete number for LLaMA 3 8B training can be found [here](https://github.com/pytorch/ao/tree/main/torchao/float8#training-benchmarks)
- Known issues:
-  - FP8 + DDP + `torch.compile` (causes [error](https://gist.github.com/djsaunde/0c1664c32e44a64d31b5e01b4aafe5c4))
-  - FP8 + FSDP2 + `torch.compile` + FSDP2 activation checkpointing tends to be _slower_ than the BF16 equivalent training
-  - Flash Attention 2 does not play nicely with `torch.compile`
-
-See `examples/llama-3/3b-fp8-fsdp2.yaml` for an optimized example config. Enabling FP8 mixed precision + FP8 all-gather training results in ~10% faster iterations per second vs. BF16 for a relatively small (3B param) model
-
-For more information on multi-GPU training, see our [Multi-GPU guide](multi-gpu.qmd).
--- a/docs/multi-gpu.qmd
+++ b/docs/multi-gpu.qmd
@@ -1,191 +0,0 @@
---
-title: "Multi-GPU"
-format:
-  html:
-    toc: true
-    toc-depth: 3
-    number-sections: true
-    code-tools: true
-execute:
-  enabled: false
---
-
-This guide covers advanced training configurations for multi-GPU setups using Axolotl.
-
-## Overview {#sec-overview}
-
-Axolotl supports several methods for multi-GPU training:
-
- DeepSpeed (recommended)
- FSDP (Fully Sharded Data Parallel)
- Sequence parallelism
- FSDP + QLoRA
-
-## DeepSpeed {#sec-deepspeed}
-
-### Configuration {#sec-deepspeed-config}
-
-Add to your YAML config:
-
-```{.yaml}
-deepspeed: deepspeed_configs/zero1.json
-```
-### Usage {#sec-deepspeed-usage}
-
-```{.bash}
-# Fetch deepspeed configs (if not already present)
-axolotl fetch deepspeed_configs
-
-# Passing arg via config
-axolotl train config.yml
-
-# Passing arg via cli
-axolotl train config.yml --deepspeed deepspeed_configs/zero1.json
-```
-
-### ZeRO Stages {#sec-zero-stages}
-
-We provide default configurations for:
-
- ZeRO Stage 1 (`zero1.json`)
- ZeRO Stage 1 with torch compile (`zero1_torch_compile.json`)
- ZeRO Stage 2 (`zero2.json`)
- ZeRO Stage 3 (`zero3.json`)
- ZeRO Stage 3 with bf16 (`zero3_bf16.json`)
- ZeRO Stage 3 with bf16 and CPU offload params(`zero3_bf16_cpuoffload_params.json`)
- ZeRO Stage 3 with bf16 and CPU offload params and optimizer (`zero3_bf16_cpuoffload_all.json`)
-
-::: {.callout-tip}
-
-Choose the configuration that offloads the least amount to memory while still being able to fit on VRAM for best performance.
-
-Start from Stage 1 -> Stage 2 -> Stage 3.
-
-:::
-
-::: {.callout-tip}
-
-Using ZeRO Stage 3 with Single-GPU training
-
-ZeRO Stage 3 can be used for training on a single GPU by manually setting the environment variables:
-`WORLD_SIZE=1 LOCAL_RANK=0 MASTER_ADDR=0.0.0.0 MASTER_PORT=29500`
-
-:::
-
-## Fully Sharded Data Parallel (FSDP) {#sec-fsdp}
-
-::: {.callout-note}
-
-FSDP2 is recommended for new users. FSDP1 is deprecated and will be removed in an upcoming release of Axolotl.
-
-:::
-
-### Migrating from FSDP1 to FSDP2 {#sec-migrate-fsdp1-fsdp2}
-
-To migrate your config from FSDP1 to FSDP2, you must use the `fsdp_version` top-level config field to specify the FSDP version, and
-also follow the config field mapping below to update field names.
-
-#### Config mapping
-
-FSDP1 | FSDP2
-------- | --------
-fsdp_sharding_strategy | reshard_after_forward
-fsdp_backward_prefetch_policy | **REMOVED**
-fsdp_backward_prefetch | **REMOVED**
-fsdp_forward_prefetch | **REMOVED**
-fsdp_sync_module_states | **REMOVED**
-fsdp_cpu_ram_efficient_loading | cpu_ram_efficient_loading
-fsdp_state_dict_type | state_dict_type
-fsdp_use_orig_params | **REMOVED**
-
-For more details, please see the migration guide in the [torchtitan repo](https://github.com/pytorch/torchtitan/blob/main/docs/fsdp.md). In Axolotl,
-if you were using the following FSDP1 config:
-
-```{.yaml}
-fsdp_version: 1
-fsdp_config:
-  fsdp_offload_params: false
-  fsdp_cpu_ram_efficient_loading: true
-  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  fsdp_transformer_layer_cls_to_wrap: Qwen3DecoderLayer
-  fsdp_state_dict_type: FULL_STATE_DICT
-  fsdp_sharding_strategy: FULL_SHARD
-```
-
-You can migrate to the following FSDP2 config:
-
-```{.yaml}
-fsdp_version: 2
-fsdp_config:
-  offload_params: false
-  cpu_ram_efficient_loading: true
-  auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  transformer_layer_cls_to_wrap: Qwen3DecoderLayer
-  state_dict_type: FULL_STATE_DICT
-  reshard_after_forward: true
-```
-
-### FSDP1 (deprecated) {#sec-fsdp-config}
-
-::: {.callout-note}
-
-Using `fsdp` to configure FSDP is deprecated and will be removed in an upcoming release of Axolotl. Please use `fsdp_config` as above instead.
-
-:::
-
-```{.yaml}
-fsdp:
-  - full_shard
-  - auto_wrap
-fsdp_config:
-  fsdp_offload_params: true
-  fsdp_state_dict_type: FULL_STATE_DICT
-  fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer
-```
-
-
-## Sequence parallelism {#sec-sequence-parallelism}
-
-We support sequence parallelism (SP) via the
-[ring-flash-attention](https://github.com/zhuzilin/ring-flash-attention) project. This
-allows one to split up sequences across GPUs, which is useful in the event that a
-single sequence causes OOM errors during model training.
-
-See our [dedicated guide](sequence_parallelism.qmd) for more information.
-
-### FSDP + QLoRA {#sec-fsdp-qlora}
-
-For combining FSDP with QLoRA, see our [dedicated guide](fsdp_qlora.qmd).
-
-## Performance Optimization {#sec-performance}
-
-### Liger Kernel Integration {#sec-liger}
-
-Please see [docs](custom_integrations.qmd#liger) for more info.
-
-## Troubleshooting {#sec-troubleshooting}
-
-### NCCL Issues {#sec-nccl}
-
-For NCCL-related problems, see our [NCCL troubleshooting guide](nccl.qmd).
-
-### Common Problems {#sec-common-problems}
-
-::: {.panel-tabset}
-
-## Memory Issues
-
- Reduce `micro_batch_size`
- Reduce `eval_batch_size`
- Adjust `gradient_accumulation_steps`
- Consider using a higher ZeRO stage
-
-## Training Instability
-
- Start with DeepSpeed ZeRO-2
- Monitor loss values
- Check learning rates
-
-:::
-
-For more detailed troubleshooting, see our [debugging guide](debugging.qmd).
--- a/docs/multi-node.qmd
+++ b/docs/multi-node.qmd
@@ -3,18 +3,6 @@ title: Multi Node
 description: How to use Axolotl on multiple machines
 ---

-The below are three ways to train multi-node in Axolotl.
-
-::: {.callout-important}
-Each machine needs a copy of Axolotl, we suggest using the same commit to ensure compatibility.
-
-You will also need to have the same configuration file for your model on each machine.
-
-Make sure the main machine is reachable by other machines.
-:::
-
-## Accelerate
-
 You will need to create a configuration for accelerate, either by using `accelerate config` and follow the instructions or you can use one of the preset below:

 ~/.cache/huggingface/accelerate/default_config.yaml
@@ -38,57 +26,23 @@ tpu_use_sudo: false
 use_cpu: false
 ```

-Configure your model to use FSDP in the Axolotl yaml. For example:
+Configure your model to use FSDP with for example:
 ```yaml
-fsdp_version: 2
+fsdp:
+  - full_shard
+  - auto_wrap
 fsdp_config:
-  offload_params: true
-  state_dict_type: FULL_STATE_DICT
-  auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  transformer_layer_cls_to_wrap: LlamaDecoderLayer
-  reshard_after_forward: true
+  fsdp_offload_params: true
+  fsdp_state_dict_type: FULL_STATE_DICT
+  fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer
 ```

+## Machine configuration
+
+On each machine you need a copy of Axolotl, we suggest using the same commit to ensure compatibility.
+
+You will also need to have the same configuration file for your model on each machine.
+
+On the main machine only, make sure the port you set as `main_process_port` is open in TCP and reachable by other machines.
+
 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.
-
-## Raytrain
-
-Please see ray train doc [here](ray-integration.qmd).
-
-## Torchrun
-
-If you are using Infiniband, we recommend torchrun to utilize the full bandwidth.
-
-Set the following env (change buffersize/socketname depending on your system):
-
-```bash
-export NCCL_IB_DISABLE=0
-export NCCL_SOCKET_IFNAME="eth0,en,eth,em,bond"
-export NCCL_BUFFSIZE=2097152
-```
-
-Run the following on each node:
-
-### Option 1: New Axolotl CLI with launcher args (Recommended)
-
-```bash
-axolotl train config.yaml --launcher torchrun -- --nnodes $num_nodes --nproc_per_node $gpu_per_node --rdzv_id $rdzv_id --rdzv_backend c10d --rdzv_endpoint "$head_node_ip:$head_node_port"
-```
-
-### Option 2: Direct torchrun (Legacy)
-
-```bash
-torchrun --nnodes $num_nodes --nproc_per_node $gpu_per_node --rdzv_id $rdzv_id --rdzv_backend c10d --rdzv_endpoint "$head_node_ip:$head_node_port" -m axolotl.cli.train config.yaml
-```
-
-Please make sure to substitute the placeholder variables:
-
- `num_nodes`: Number of nodes (containing GPUs)
- `gpu_per_node`: Number of gpus per node
- `head_node_ip`: IP of the head node (make sure other machines can connect to this)
- `head_node_port`: Port of the head node (make sure other machines can connect to this. Default 29400)
- `rdzv_id`: A unique job ID that is used by the job across nodes.
-
-The new CLI approach (Option 1) is recommended as it provides consistent argument handling and works seamlessly with other Axolotl CLI features.
-
-More info on the available configs can be found on the Pytorch docs [here](https://pytorch.org/docs/stable/elastic/run.html)
--- a/docs/multimodal.qmd
+++ b/docs/multimodal.qmd
@@ -1,266 +1,28 @@
---
-title: MultiModal / Vision Language Models (BETA)
-format:
-  html:
-    toc: true
-    toc-depth: 3
---
+# MultiModal / Vision Language Models (BETA)

-## Supported Models
+### Supported Models

- [Mllama](#sec-mllama)
- [Llama4](#sec-llama4)
- [Pixtral](#sec-pixtral)
- [Llava-1.5](#sec-llava-15)
- [Mistral-Small-3.1](#sec-mistral-small-31)
- [Voxtral](#sec-voxtral)
- [Gemma-3](#sec-gemma-3)
- [Gemma-3n](#sec-gemma-3n)
- [Qwen2-VL](#sec-qwen2-vl)
- [Qwen2.5-VL](#sec-qwen25-vl)
- [SmolVLM2](#sec-smolvlm2)
- [LFM2-VL](#sec-lfm2-vl)
+- Mllama, i.e. llama with vision models

-## Usage
+### Usage

-Multimodal support is limited and doesn't have full feature parity.
-
-Here are the hyperparams you'll need to use to finetune a multimodal model.
+Currently multimodal support is limited and doesn't have full feature parity. To finetune a multimodal Llama w/ LoRA,
+you'll need to use the following in YAML in combination with the rest of the required hyperparams.

 ```yaml
+base_model: alpindale/Llama-3.2-11B-Vision-Instruct
 processor_type: AutoProcessor
-
 skip_prepare_dataset: true
-remove_unused_columns: false  # leave columns in place as they are needed to handle image embeddings during training
-sample_packing: false  # not yet supported with multimodal

-chat_template:  # see in next section if specified
-
-# example dataset
+chat_template: llama3_2_vision
 datasets:
  - path: HuggingFaceH4/llava-instruct-mix-vsft
    type: chat_template
    split: train[:1%]
    field_messages: messages
+remove_unused_columns: false
+sample_packing: false

-# (optional) if doing lora, only finetune the Language model,
-# leave the vision model and vision tower frozen
-# load_in_8bit: true
-adapter: lora
-lora_target_modules: 'model.language_model.layers.[\d]+.(mlp|cross_attn|self_attn).(up|down|gate|q|k|v|o)_proj'
-
-# (optional) if you want to resize images to a set size
-image_size: 512
-image_resize_algorithm: bilinear
+# only finetune the Language model, leave the vision model and vision tower frozen
+lora_target_modules: 'language_model.model.layers.[\d]+.(mlp|cross_attn|self_attn).(up|down|gate|q|k|v|o)_proj'
 ```
-
-Please see [examples](https://github.com/axolotl-ai/axolotl/tree/main/examples) folder for full configs.
-
-::: {.callout-warning}
-Some of our chat_templates have been extended to support broader dataset types. This should not break any existing configs.
-:::
-
-### Mllama {#sec-mllama}
-
-```yaml
-base_model: meta-llama/Llama-3.2-11B-Vision-Instruct
-
-chat_template: llama3_2_vision
-```
-
-### Llama4 {#sec-llama4}
-
-```yaml
-base_model: meta-llama/Llama-4-Scout-17B-16E-Instruct
-
-chat_template: llama4
-```
-
-### Pixtral {#sec-pixtral}
-
-```yaml
-base_model: mistralai/Pixtral-12B-2409
-
-chat_template: pixtral
-```
-
-### Llava-1.5 {#sec-llava-15}
-
-```yaml
-base_model: llava-hf/llava-1.5-7b-hf
-
-chat_template: llava
-```
-
-### Mistral-Small-3.1 {#sec-mistral-small-31}
-
-```yaml
-base_model: mistralai/Mistral-Small-3.1-24B-Instruct-2503
-
-chat_template: mistral_v7_tekken
-```
-
-### Voxtral {#sec-voxtral}
-
-::: {.callout-tip}
-Please make sure to install audio lib via `pip3 install librosa==0.11.0 'mistral_common[audio]==1.8.3'`
-:::
-
-```yaml
-base_model: mistralai/Voxtral-Mini-3B-2507
-```
-
-### Gemma-3 {#sec-gemma-3}
-
-::: {.callout-tip}
-The Gemma3-1B model is a text-only model, so please train as regular text model.
-:::
-
-For multi-modal 4B/12B/27B models, use the following config:
-
-```yaml
-base_model: google/gemma-3-4b-it
-
-chat_template: gemma3
-```
-
-### Gemma-3n {#sec-gemma-3n}
-
-::: {.callout-warning}
-The model's initial loss and grad norm will be very high. We suspect this to be due to the Conv in the vision layers.
-:::
-
-::: {.callout-tip}
-Please make sure to install `timm` via `pip3 install timm==1.0.17`
-:::
-
-```yaml
-base_model: google/gemma-3n-E2B-it
-
-chat_template: gemma3n
-```
-
-### Qwen2-VL {#sec-qwen2-vl}
-
-```yaml
-base_model: Qwen/Qwen2-VL-7B-Instruct
-
-chat_template: qwen2_vl
-```
-
-### Qwen2.5-VL {#sec-qwen25-vl}
-
-```yaml
-base_model: Qwen/Qwen2.5-VL-7B-Instruct
-
-chat_template: qwen2_vl  # same as qwen2-vl
-```
-
-### SmolVLM2 {#sec-smolvlm2}
-
-::: {.callout-tip}
-Please make sure to install `num2words` via `pip3 install num2words==0.5.14`
-:::
-
-```yaml
-base_model: HuggingFaceTB/SmolVLM2-500M-Video-Instruct
-```
-
-### LFM2-VL {#sec-lfm2-vl}
-
-::: {.callout-warning}
-Please uninstall `causal-conv1d` via `pip3 uninstall -y causal-conv1d`
-:::
-
-```yaml
-base_model: LiquidAI/LFM2-VL-450M
-```
-
-## Dataset Format
-
-For multi-modal datasets, we adopt an extended `chat_template` format similar to OpenAI's Message format.
-
- A message is a list of `role` and `content`.
- `role` can be `system`, `user`, `assistant`, etc.
- `content` is a list of `type` and (`text`, `image`, `path`, `url`, `base64`, or `audio`).
-
-### Image
-
-::: {.callout-note}
-For backwards compatibility:
-
- If the dataset has a `images` or `image` column of `list[Image]`, it will be appended to the first `content` list as `{"type": "image", "image": ...}`. However, if the content already has a `{"type": "image"}` but no `image` key, it will be set the `image` key.
- If `content` is a string, it will be converted to a list with `type` as `text`.
-:::
-
-For image loading, you can use the following keys within `content` alongside `"type": "image"`:
-
- `"path": "/path/to/image.jpg"`
- `"url": "https://example.com/image.jpg"`
- `"base64": "..."`
- `"image": PIL.Image`
-
-### Audio
-
-For audio loading, you can use the following keys within `content` alongside `"type": "audio"`:
-
- `"path": "/path/to/audio.mp3"`
- `"url": "https://example.com/audio.mp3"`
- `"audio": np.ndarray`
-
-::: {.callout-tip}
-
-You may need to install `librosa` via `pip3 install librosa==0.11.0`.
-
-:::
-
-### Video
-
-::: {.callout-warning}
-
-This is not well tested at the moment. We welcome contributors!
-
-:::
-
-For video loading, you can use the following keys within `content` alongside `"type": "video"`:
-
- `"path": "/path/to/video.mp4"`
- `"url": "https://example.com/video.mp4"`
- `"video": np.ndarray | list[PIL.Image.Image] | torch.Tensor` (or list of the aforementioned)
-
-### Example
-
-Here is an example of a multi-modal dataset:
-```json
-[
-  {
-    "messages": [
-        {
-            "role": "system",
-            "content": [
-              {"type": "text", "text": "You are a helpful assistant."}
-              ]
-        },
-        {
-            "role": "user",
-            "content": [
-                {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
-                {"type": "text", "text": "Describe this image in detail."}
-            ]
-        },
-        {
-            "role": "assistant",
-            "content": [
-              {"type": "text", "text": "The image is a bee."}
-            ]
-        }
-    ]
-  }
-]
-```
-
-## FAQ
-
-1. `PIL.UnidentifiedImageError: cannot identify image file ...`
-
-`PIL` could not retrieve the file at `url` using `requests`. Please check for typo. One alternative reason is that the request is blocked by the server.
--- a/docs/nccl.qmd
+++ b/docs/nccl.qmd
@@ -13,13 +13,13 @@ Often, this timeout will happen after 30 minutes (the default setting) and is ac

 Forcing cross-GPU communication via [NVLink](https://en.wikipedia.org/wiki/NVLink) may help without increasing timeouts. To verify that your configuration is leveraging NVLink run the following command:

-```bash
+```shell
 nvidia-smi nvlink --status
 ```

 To force NCCL to use NVLink, simply set this in the environment:

-```bash
+```shell
 export NCCL_P2P_LEVEL=NVL
 ```

@@ -33,13 +33,13 @@ If NVLink is not available in your environment there are other options for ``NCC

 To validate that acceptable data transfer speeds exist for your training job, running [NCCL Tests](https://github.com/NVIDIA/nccl-tests/blob/master/README.md) can help pinpoint bottlenecks, for example:

-```bash
+```shell
 ./build/all_reduce_perf -b 8 -e 128M -f 2 -g 3
 ```

 It can be useful when debugging NCCL communication timeouts to activate additional logging in both PyTorch and NCCL:

-```bash
+```shell
 export NCCL_DEBUG=INFO
 export NCCL_DEBUG_SUBSYS=ALL
 export TORCH_DISTRIBUTED_DEBUG=INFO
--- a/docs/nd_parallelism.qmd
+++ b/docs/nd_parallelism.qmd
@@ -1,108 +0,0 @@
---
-title: "N-D Parallelism (Beta)"
---
-
-Axolotl enables training models at scale by composing different parallelism techniques. This is essential when:
-
- A model's weights are too large to fit on a single GPU's memory.
- A model's activations, especially with very long contexts, are too large for a single GPU.
- You want to accelerate training by using multiple GPUs or nodes.
-
-or combinations of the above!
-
-## Core Concepts
-
-Parallelism strategies can be combined. The key is understanding how each one divides the workload. PyTorch's `DeviceMesh` is the modern way to manage these combinations, creating a logical grid of your GPUs and assigning different parallel strategies to different dimensions of the grid.
-
-### Data Parallelism {#sec-dp}
-
-Data Parallelism focuses on splitting the global data batch across GPUs.
-
- Distributed Data Parallel (DDP): The classic approach. The full model is replicated on every GPU. Each GPU processes a different slice of the data batch. Gradients are then averaged across all GPUs after the backward pass to keep the models synchronized. This can substantially improve data throughput compared to single-device training, but requires that each GPU is able to hold the entire model, its gradients, and optimizer states.
-
- [Fully Sharded Data Parallel (FSDP)](multi-gpu.qmd#fully-sharded-data-parallel-(fsdp)): A highly memory-efficient form of data parallelism (inspired by DeepSpeed's ZeRO). Instead of replicating the model, FSDP shards the model's *parameters, gradients, and optimizer states* across the GPUs in the data-parallel group. During computation, each GPU receives the specific parameters it needs via an `all_gather` operation just before they are used, and they can be discarded immediately after (`reshard-after-forward`).
-    - FSDP maps to ZeRO stages:
-        - ZeRO-2 (`reshard_after_forward=False`): Shards gradients and optimizer states. Model weights are replicated on each GPU.
-        - ZeRO-3 (`reshard_after_forward=True`): Shards gradients, optimizer states, AND model parameters. This provides the most memory savings at the cost of more communication (re-gathering parameters for both forward and backward passes).
-
-### [Experimental] Tensor Parallelism (TP) {#sec-tp}
-
-Also known as "horizontal model parallelism," as described in the [Megatron-LM paper](https://arxiv.org/pdf/1909.08053.pdf). Instead of splitting the batch, TP splits the model's layers themselves across GPUs.
-
- How it works: For a linear layer `Y = XA`, the weight matrix `A` is split column-wise (`A = [A_1, A_2]`). The computation becomes `Y_1 = XA_1` and `Y_2 = XA_2`, which can happen in parallel on different GPUs. The final output `Y` is simply the concatenation of `Y_1` and `Y_2`. Check [this comment](https://github.com/huggingface/transformers/issues/10321#issuecomment-783543530) for more detailed info.
- Requirement: TP involves frequent, small communications within a forward/backward pass. It requires a very fast interconnect between GPUs (e.g., NVLink) and is typically not recommended across different nodes.
-
-### Context Parallelism (CP) {#sec-cp}
-
-Context Parallelism, also called [Sequence Parallelism](sequence_parallelism.qmd), addresses the memory bottleneck from long sequences. The input sequence itself is split along the sequence length dimension and distributed across GPUs.
-
- How it works: If you have a sequence of 8192 tokens and a `context_parallel_size` of 4, each GPU will only handle a chunk of 2048 tokens.
- The Challenge: Attention is not local; every token needs to "attend to" every other token. Splitting the sequence breaks this.
- The Solution (`ring-flash-attention`): An efficient communication protocol is used. To compute attention for its local sequence chunk, each GPU passes its Key-Value (KV) cache to its neighbor in a "ring." After `N-1` steps, every GPU has seen the KV-cache from all other GPUs, allowing it to compute the correct attention values for its chunk. This is implemented using the highly optimized `flash-attention` kernel at each step.
-
-### Hybrid Sharding Data Parallel (HSDP) {#sec-hsdp}
-
-HSDP is a 2D strategy that intelligently combines FSDP and DDP, typically for multi-node training.
-
- Intra-Node (within a machine): Use FSDP. This is efficient because GPUs on the same node have fast interconnects (NVLink), making the `all_gather` operations for sharded parameters fast.
- Inter-Node (across machines): Use DDP. The gradient synchronization between nodes is less frequent than FSDP's parameter gathering, making it a better fit for the slower node-to-node network (e.g., Ethernet/Infiniband).
- Example: With 2 nodes of 8 GPUs each (16 total), you could have `dp_shard_size=8` (FSDP within each node) and `dp_replicate_size=2` (DDP across the two nodes).
-
-## Usage
-
-```yaml
-# FSDP config. See https://docs.axolotl.ai/docs/multi-gpu.html#sec-fsdp
-fsdp_version: 2
-fsdp_config:
-  # ...
-
-# The number of GPUs to shard the model parameters across (FSDP dimension).
-dp_shard_size: 4
-
-# The number of times to replicate the sharded model (DDP dimension).
-dp_replicate_size: 2
-
-# Number of GPUs for Tensor Parallelism.
-tensor_parallel_size: 1  # (default is 1, no TP)
-
-# Number of GPUs for Context/Sequence Parallelism.
-context_parallel_size: 1 # (default is 1, no CP)
-```
-
-Note: We recommend FSDP. DeepSpeed is only compatible with `tensor_parallel_size`.
-
-## Examples
-
-::: {.callout-tip}
-See our example configs [here](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/distributed-parallel).
-:::
-
-1.  HSDP on 2 nodes with 4 GPUs each (8 GPUs total):
-    - You want FSDP within each node and DDP across nodes.
-    - Set `dp_shard_size: 4` and `dp_replicate_size: 2`.
-
-2.  FSDP + TP on a single 8-GPU node:
-    - You want to split the model across 4 GPUs using FSDP, and further split each layer across 2 GPUs with TP.
-    - Set `dp_shard_size: 4` and `tensor_parallel_size: 2`.
-
-3.  FSDP + CP on a single 8-GPU node for long context:
-    - You want to shard the model across all 8 GPUs and also split the sequence length across all 8 GPUs.
-    - Set `dp_shard_size: 8` and `context_parallel_size: 8`. Note: this means the data parallel group and context parallel group are the same. A more common setup might be to shard across a smaller group.
-
-## Support Matrix
-
-This matrix describes how different parallelism methods can be combined in Axolotl.
-
-| Combination | `dp_replicate_size` | `dp_shard_size` | `tp_size` | `cp_size` | Status & Notes |
-| --- | :---: | :---: |:---:|:---:|---|
-| **FSDP** (ZeRO-3) | 1 | >1 | 1 | 1 | ✅ Fully supported. Shards model across all GPUs. |
-| **HSDP** | >1 | >1 | 1 | 1 | ✅ Fully supported. FSDP intra-node, DDP inter-node. |
-| **FSDP + TP** | 1 | >1 | >1 | 1 | ✅ **2D Parallelism**. Shards the model across a `dp_shard` group, and TP-splits layers within the `tp` group. |
-| **HSDP + TP** | >1 | >1 | >1 | 1 | ✅ **3D Parallelism**. A powerful but complex combination. |
-| **FSDP + CP** | 1 | >1 | 1 | >1 | ✅ **2D Parallelism**. Combines FSDP with context parallelism. |
-| **FSDP + TP + CP**| 1 | >1 | >1| >1| ✅ **3D Parallelism**. Another advanced combination. |
-| DDP + TP/CP | >1 | 1 | >1 | >1 | ❌ **Not Supported**. The `ParallelismConfig` explicitly prevents this, as composing pure DDP with TP or CP is currently not supported. You should use FSDP + TP/CP instead (`dp_shard_size > 1`). |
-| Just TP / CP | 1 | 1 | >1 | >1 | ✅ Supported. Useful for inference or when the model fits on one GPU but context is too long. |
-
- `tp_size` refers to `tensor_parallel_size`
- `cp_size` refers to `context_parallel_size`
--- a/docs/optimizers.qmd
+++ b/docs/optimizers.qmd
@@ -1,129 +0,0 @@
---
-title: Optimizers
-description: Configuring optimizers
---
-
-## Overview
-
-Axolotl supports all optimizers supported by [transformers OptimizerNames](https://github.com/huggingface/transformers/blob/51f94ea06d19a6308c61bbb4dc97c40aabd12bad/src/transformers/training_args.py#L142-L187)
-
-Here is a list of optimizers supported by transformers as of `v4.54.0`:
-
- `adamw_torch`
- `adamw_torch_fused`
- `adamw_torch_xla`
- `adamw_torch_npu_fused`
- `adamw_apex_fused`
- `adafactor`
- `adamw_anyprecision`
- `adamw_torch_4bit`
- `adamw_torch_8bit`
- `ademamix`
- `sgd`
- `adagrad`
- `adamw_bnb_8bit`
- `adamw_8bit`  # alias for adamw_bnb_8bit
- `ademamix_8bit`
- `lion_8bit`
- `lion_32bit`
- `paged_adamw_32bit`
- `paged_adamw_8bit`
- `paged_ademamix_32bit`
- `paged_ademamix_8bit`
- `paged_lion_32bit`
- `paged_lion_8bit`
- `rmsprop`
- `rmsprop_bnb`
- `rmsprop_bnb_8bit`
- `rmsprop_bnb_32bit`
- `galore_adamw`
- `galore_adamw_8bit`
- `galore_adafactor`
- `galore_adamw_layerwise`
- `galore_adamw_8bit_layerwise`
- `galore_adafactor_layerwise`
- `lomo`
- `adalomo`
- `grokadamw`
- `schedule_free_radam`
- `schedule_free_adamw`
- `schedule_free_sgd`
- `apollo_adamw`
- `apollo_adamw_layerwise`
- `stable_adamw`
-
-
-## Custom Optimizers
-
-Enable custom optimizers by passing a string to the `optimizer` argument. Each optimizer will receive beta and epsilon args, however, some may accept additional args which are detailed below.
-
-### optimi_adamw
-
-```yaml
-optimizer: optimi_adamw
-```
-
-### ao_adamw_4bit
-
-Deprecated: Please use `adamw_torch_4bit`.
-
-### ao_adamw_8bit
-
-Deprecated: Please use `adamw_torch_8bit`.
-
-### ao_adamw_fp8
-
-
-```yaml
-optimizer: ao_adamw_fp8
-```
-
-### adopt_adamw
-
-GitHub: [https://github.com/iShohei220/adopt](https://github.com/iShohei220/adopt)
-Paper: [https://arxiv.org/abs/2411.02853](https://arxiv.org/abs/2411.02853)
-
-```yaml
-optimizer: adopt_adamw
-```
-
-### came_pytorch
-
-GitHub: [https://github.com/yangluo7/CAME/tree/master](https://github.com/yangluo7/CAME/tree/master)
-Paper: [https://arxiv.org/abs/2307.02047](https://arxiv.org/abs/2307.02047)
-
-```yaml
-optimizer: came_pytorch
-
-# optional args (defaults below)
-adam_beta1: 0.9
-adam_beta2: 0.999
-adam_beta3: 0.9999
-adam_epsilon: 1e-30
-adam_epsilon2: 1e-16
-```
-
-### muon
-
-Blog: [https://kellerjordan.github.io/posts/muon/](https://kellerjordan.github.io/posts/muon/)
-Paper: [https://arxiv.org/abs/2502.16982v1](https://arxiv.org/abs/2502.16982v1)
-
-```yaml
-optimizer: muon
-```
-
-### dion
-
-Microsoft's Dion (DIstributed OrthoNormalization) optimizer is a scalable and communication-efficient
-orthonormalizing optimizer that uses low-rank approximations to reduce gradient communication.
-
-GitHub: [https://github.com/microsoft/dion](https://github.com/microsoft/dion)
-Paper: [https://arxiv.org/pdf/2504.05295](https://arxiv.org/pdf/2504.05295)
-Note: Implementation written for PyTorch 2.7+ for DTensor
-
-```yaml
-optimizer: dion
-dion_lr: 0.01
-dion_momentum: 0.95
-lr: 0.00001  # learning rate for embeddings and parameters that fallback to AdamW
-```
--- a/docs/qat.qmd
+++ b/docs/qat.qmd
@@ -1,32 +0,0 @@
---
-title: "Quantization Aware Training (QAT)"
-back-to-top-navigation: true
-toc: true
-toc-expand: 2
-toc-depth: 4
---
-
-## Overview
-
-[Quantization Aware Training](https://pytorch.org/blog/introduction-to-quantization-on-pytorch/#quantization-aware-training) (QAT) is a technique for improving the accuracy of models which are quantized
-by applying "fake" quantizations to the model's weights (and optionally, activations) during training. This fake
-quantization allows for the model to adjust for noise introduced by the quantization, so when the model is eventually
-quantized, the accuracy loss is minimized. We use the quantization techniques implemented in [torchao](https://github.com/pytorch/ao) to provide
-support for QAT and post-training quantization (PTQ) in axolotl.
-
-We recommend reviewing the excellent QAT tutorial in the [torchtune library](https://pytorch.org/torchtune/main/tutorials/qat_finetune.html#quantizing-the-qat-model),
-and the QAT documentation in the [torchao library](https://github.com/pytorch/ao/tree/main/torchao/quantization/qat), for more details.
-
-## Configuring QAT in Axolotl
-
-To enable QAT in axolotl, add the following to your configuration file:
-
-```yaml
-qat:
-  activation_dtype: # Optional[str] = "int8". Fake quantization layout to use for activation quantization. Valid options are "int4" and "int8"
-  weight_dtype: # Optional[str] = "int8". Fake quantization layout to use for weight quantization. Valid options are "int4" and "int8"
-  group_size: # Optional[int] = 32. The number of elements in each group for per-group fake quantization
-  fake_quant_after_n_steps: # Optional[int] = None. The number of steps to apply fake quantization after
-```
-
-Once you have finished training, you must quantize your model by using the same quantization configuration which you used to train the model with. You can use the [`quantize`](./quantize.qmd) command to do this.
--- a/docs/quantize.qmd
+++ b/docs/quantize.qmd
@@ -1,53 +0,0 @@
---
-title: "Quantization with torchao"
-back-to-top-navigation: true
-toc: true
-toc-expand: 2
-toc-depth: 4
---
-
-Quantization is a technique to lower the memory footprint of your model, potentially at the cost of accuracy or model performance. We support quantizing your model using the [torchao](https://github.com/pytorch/ao) library. Quantization is supported for both post-training quantization (PTQ) and quantization-aware training (QAT).
-
-
-::: {.callout-note}
-
-We do not currently support quantization techniques such as GGUF/GPTQ,EXL2 at the moment.
-
-:::
-
-## Configuring Quantization in Axolotl
-
-Quantization is configured using the `quantization` key in your configuration file.
-
-```yaml
-base_model: # The path to the model to quantize.
-quantization:
-  weight_dtype: # Optional[str] = "int8". Fake quantization layout to use for weight quantization. Valid options are uintX for X in [1, 2, 3, 4, 5, 6, 7], or int4, or int8
-  activation_dtype: # Optional[str] = "int8". Fake quantization layout to use for activation quantization. Valid options are "int4" and "int8"
-  group_size: # Optional[int] = 32. The number of elements in each group for per-group fake quantization
-  quantize_embedding: # Optional[bool] = False. Whether to quantize the embedding layer.
-
-output_dir:  # The path to the output directory.
-```
-
-Once quantization is complete, your quantized model will be saved in the `{output_dir}/quantized` directory.
-
-You may also use the `quantize` command to quantize a model which has been trained with [QAT](./qat.qmd) - you can do this by using the existing QAT configuration file which
-you used to train the model:
-
-```yaml
-# qat.yml
-qat:
-  activation_dtype: int8
-  weight_dtype: int8
-  group_size: 256
-  quantize_embedding: true
-
-output_dir: # The path to the output directory used during training where the final checkpoint has been saved.
-```
-
-```bash
-axolotl quantize qat.yml
-```
-
-This ensures that an identical quantization configuration is used to quantize the model as was used to train it.
--- a/docs/ray-integration.qmd
+++ b/docs/ray-integration.qmd
@@ -1,91 +0,0 @@
---
-title: Ray Train
-description: How to use Axolotl with Ray Train
---
-
-Axolotl supports using Ray as an alternative to `accelerate` for orchestrating training. This is especially useful for multi-node training since you only have to setup code and dependencies in a single node and launch training as if you were using a single node.
-
-With the `--use-ray` CLI flag, Axolotl will use Ray Train's [`TorchTrainer`](https://docs.ray.io/en/latest/train/api/doc/ray.train.torch.TorchTrainer.html#ray.train.torch.TorchTrainer) to run training.
-
-## Ray cluster setup
-
-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).
-
-Every Ray cluster has one _head_ 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 [doc](https://docs.ray.io/en/latest/cluster/key-concepts.html#cluster-key-concepts).
-
-## Sanity check
-
-To run a sanity check on whether your ray cluster is setup properly, execute the following on the head node:
-
-```bash
-ray status
-```
-
-The output should have a summary of your Ray cluster - list of all the nodes in your cluster, the number of CPUs and GPUs in your cluster, etc. For example, if you have a cluster with 1 CPU-only head node and 2 4xL40S worker nodes, the output can look like this:
-
-
-```
-Node status
---------------------------------------------------------------
-Active:
- 1 head
-Idle:
- 2 4xL40S:48CPU-384GB
-Pending:
- (no pending nodes)
-Recent failures:
- (no failures)
-
-Resources
---------------------------------------------------------------
-Usage:
- 0.0/96.0 CPU
- 0.0/8.0 GPU
- 0B/800.00GiB memory
- 0B/229.57GiB object_store_memory
-
-Demands:
- (no resource demands)
-```
-
-You should also be able to see the same on the [Ray dashboard](https://docs.ray.io/en/latest/ray-observability/getting-started.html).
-
-
-## Configuring training with Ray Train
-
-You can find an example configuration at `configs/llama-3/lora-1b-ray.yaml`.
-
-The key parameters to note here are:
-
-```yaml
-use_ray: true
-ray_num_workers: 4
-# optional
-resources_per_worker:
-    GPU: 1
-```
-
- `use_ray`: This is the flag that enables the Ray Train integration. You can either use the corresponding `--use-ray` flag in the CLI or set `use_ray` in the config file.
- `ray_num_workers`: This is the number of workers/GPUs to use for training.
- `resources_per_worker`: This is the Ray [resource request](https://docs.ray.io/en/latest/ray-core/scheduling/resources.html) for each worker. This can be used to request a specific GPU type or a custom resource for each worker. For example, if your ray cluster has GPUs of different types, and you only want to use NVIDIA L40S GPUs, you can do
-
-```yaml
-resources_per_worker:
-    accelerator_type:L40S: 0.001
-```
-
-## Launching training
-
-You can simply run the following command on the head node:
-
-```bash
-axolotl train examples/llama-3/lora-1b-ray.yml --use-ray
-```
-
-This will launch training on the head node and workers will be scheduled automatically by Ray Train to run on the appropriate head or worker nodes.
-
-You can also monitor training progress on the Ray dashboard.
-
-Coming back to the example on a Ray cluster with 1 head node and 2 4xL40S worker nodes, let's say you want to make use of all 8 GPUs. You would be able to just set `ray_num_workers: 8` and run the previous command. The Cluster tab will show the following:
-
-![Ray dashboard](./images/ray-cluster-dashboard.png)
--- a/docs/reward_modelling.qmd
+++ b/docs/reward_modelling.qmd
@@ -1,64 +0,0 @@
---
-title: "Reward Modelling"
-description: "Reward models are used to guide models towards behaviors which is preferred by humans, by training over large datasets annotated with human preferences. "
---
-
-### Overview
-
-Reward modelling is a technique used to train models to predict the reward or value of a given input. This is particularly useful in reinforcement learning scenarios where the model needs to evaluate the quality of its actions or predictions.
-We support the reward modelling techniques supported by `trl`.
-
-### (Outcome) Reward Models
-
-Outcome reward models are trained using data which contains preference annotations for an entire interaction between the user and model (e.g. rather than per-turn or per-step).
-
-```yaml
-base_model: google/gemma-2-2b
-model_type: AutoModelForSequenceClassification
-num_labels: 1
-tokenizer_type: AutoTokenizer
-
-reward_model: true
-chat_template: gemma
-datasets:
-  - path: argilla/distilabel-intel-orca-dpo-pairs
-    type: bradley_terry.chat_template
-
-val_set_size: 0.1
-eval_steps: 100
-```
-
-Bradley-Terry chat templates expect single-turn conversations in the following format:
-
-```json
-{
-    "system": "...", // optional
-    "input": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-### Process Reward Models (PRM)
-
-::: {.callout-tip}
-Check out our [PRM blog](https://axolotlai.substack.com/p/process-reward-models).
-:::
-
-Process reward models are trained using data which contains preference annotations for each step in a series of interactions. Typically, PRMs are trained to provide reward signals over each step of a reasoning trace and are used for downstream reinforcement learning.
-```yaml
-base_model: Qwen/Qwen2.5-3B
-model_type: AutoModelForTokenClassification
-num_labels: 2
-
-process_reward_model: true
-datasets:
-  - path: trl-lib/math_shepherd
-    type: stepwise_supervised
-    split: train
-
-val_set_size: 0.1
-eval_steps: 100
-```
-
-Please see [stepwise_supervised](dataset-formats/stepwise_supervised.qmd) for more details on the dataset format.
--- a/docs/rlhf.qmd
+++ b/docs/rlhf.qmd
@@ -1,40 +1,26 @@
 ---
 title: "RLHF (Beta)"
 description: "Reinforcement Learning from Human Feedback is a method whereby a language model is optimized from data using human feedback."
-back-to-top-navigation: true
-toc: true
-toc-expand: 2
-toc-depth: 4
 ---

-## Overview
+### Overview

 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:

- [Direct Preference Optimization (DPO)](#dpo)
- [Identity Preference Optimization (IPO)](#ipo)
- [Kahneman-Tversky Optimization (KTO)](#kto)
- [Odds Ratio Preference Optimization (ORPO)](#orpo)
- [Group Relative Policy Optimization (GRPO)](#grpo)
+- Proximal Policy Optimization (PPO) (not yet supported in axolotl)
+- Direct Preference Optimization (DPO)
+- Identity Preference Optimization (IPO)


-## RLHF using Axolotl
+### RLHF using Axolotl

-::: {.callout-important}
-This is a BETA feature and many features are not fully implemented. You are encouraged to open new PRs to improve the integration and functionality.
-:::
+>[!IMPORTANT]
+>This is a BETA feature and many features are not fully implemented. You are encouraged to open new PRs to improve the integration and functionality.

-We rely on the [TRL](https://github.com/huggingface/trl) library for implementations of various RL training methods, which we wrap around to expose in axolotl. Each method has their own supported ways of loading datasets and prompt formats.
-
-::: {.callout-tip}
-You can find what each method supports by going into `src/axolotl/prompt_strategies/{method}` where `{method}` is one of our supported methods. The `type: ` can be retrieved from `{method}.{function_name}`.
-:::
-
-### DPO
-
-Example config:
+The various RL training methods are implemented in trl and wrapped via axolotl. Below are various examples with how you can use various preference datasets to train models that use ChatML

+#### DPO
 ```yaml
 rl: dpo
 datasets:
@@ -43,267 +29,15 @@ datasets:
    type: chatml.intel
  - path: argilla/ultrafeedback-binarized-preferences
    split: train
-    type: chatml
+    type: chatml.argilla
 ```

-DPO supports the following types with the following dataset format:
-
-#### chatml.argilla
-
-```json
-{
-    "system": "...", // optional
-    "instruction": "...",
-    "chosen_response": "...",
-    "rejected_response": "..."
-}
-```
-
-#### chatml.argilla_chat
-
-```json
-{
-    "chosen": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ],
-    "rejected": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### chatml.icr
-
-```json
-{
-    "system": "...", // optional
-    "input": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### chatml.intel
-
-```json
-{
-    "system": "...", // optional
-    "question": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### chatml.prompt_pairs
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### chatml.ultra
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "chosen": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ],
-    "rejected": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### llama3.argilla
-
-```json
-{
-    "system": "...", // optional
-    "instruction": "...",
-    "chosen_response": "...",
-    "rejected_response": "..."
-}
-```
-
-#### llama3.argilla_chat
-
-```json
-{
-    "chosen": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ],
-    "rejected": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### llama3.icr
-
-```json
-{
-    "system": "...", // optional
-    "input": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### llama3.intel
-
-```json
-{
-    "system": "...", // optional
-    "question": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### llama3.prompt_pairs
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-#### llama3.ultra
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "chosen": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ],
-    "rejected": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### zephyr.nectar
-
-```json
-{
-    "prompt": "...",
-    "answers": [
-        {
-            "answer": "...",
-            "rank": 1
-        },
-        {
-            "answer": "...",
-            "rank": 2
-        }
-        // ... more answers with ranks
-    ]
-}
-```
-
-#### chat_template.default
-
-```yaml
-rl: dpo
-datasets:
-  - path: ...
-    split: train
-    type: chat_template.default
-    field_messages: "messages"
-    field_chosen: "chosen"
-    field_rejected: "rejected"
-    message_property_mappings:
-      role: role
-      content: content
-    roles:
-      user: ["user"]
-      assistant: ["assistant"]
-      system: ["system"]
-```
-
-Sample input format:
-
-```json
-{
-    "messages": [
-        {
-            "role": "system",
-            "content": "..."
-        },
-        {
-            "role": "user",
-            "content": "..."
-        },
-        // ... more messages
-    ],
-    "chosen": {
-        "role": "assistant",
-        "content": "..."
-    },
-    "rejected": {
-        "role": "assistant",
-        "content": "..."
-    }
-}
-```
-
-#### user_defined.default
-
-For custom behaviors,
-
-```yaml
-rl: dpo
-datasets:
-  - path: ...
-    split: train
-    type:
-      field_prompt: "prompt"
-      field_system: "system"
-      field_chosen: "chosen"
-      field_rejected: "rejected"
-      prompt_format: "{prompt}"
-      chosen_format: "{chosen}"
-      rejected_format: "{rejected}"
-```
-
-The input format is a simple JSON input with customizable fields based on the above config.
-
-```json
-{
-    "system": "...",  // optional
-    "prompt": "...",
-    "chosen": "...",
-    "rejected": "..."
-}
-```
-
-### IPO
-
-As IPO is just DPO with a different loss function, all supported dataset formats for [DPO](#dpo) are also supported for IPO.
-
+#### IPO
 ```yaml
 rl: ipo
 ```

-### ORPO
+#### ORPO

 Paper: https://arxiv.org/abs/2403.07691

@@ -318,34 +52,13 @@ datasets:
    type: chat_template.argilla
 ```

-ORPO supports the following types with the following dataset format:

-#### chat_template.argilla
-
-```json
-{
-    "system": "...",  // optional
-    "prompt": "...",  // if available, will be taken as user message for single-turn instead of from list below
-
-    // chosen/rejected should be same till last content and only even-number of alternating user/assistant turns
-    "chosen": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ],
-    "rejected": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-### KTO
+#### KTO

 ```yaml
 rl: kto
-rl_beta: 0.1  # default
-kto_desirable_weight: 1.0  # default
-kto_undesirable_weight: 1.0  # default
+rl_beta: 0.5
+kto_desirable_weight: 0.2

 remove_unused_columns: false

@@ -359,257 +72,7 @@ gradient_checkpointing_kwargs:
  use_reentrant: true
 ```

-KTO supports the following types with the following dataset format:
-
-#### chatml.argilla
-
-```json
-{
-    "system": "...", // optional
-    "instruction": "...",
-    "completion": "..."
-}
-```
-
-#### chatml.argilla_chat
-
-```json
-{
-    "chosen": [
-        {"role": "user", "content": "..."}
-    ],
-    "completion": [
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### chatml.intel
-
-```json
-{
-    "system": "...", // optional
-    "question": "...",
-    "completion": "..."
-}
-```
-
-#### chatml.prompt_pairs
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "completion": "..."
-}
-```
-
-#### chatml.ultra
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "completion": "..."
-}
-```
-
-#### llama3.argilla
-
-```json
-{
-    "system": "...", // optional
-    "instruction": "...",
-    "completion": "..."
-}
-```
-
-#### llama3.argilla_chat
-
-```json
-{
-    "completion": [
-        {"role": "user", "content": "..."},
-        {"role": "assistant", "content": "..."}
-    ]
-}
-```
-
-#### llama3.intel
-
-```json
-{
-    "system": "...", // optional
-    "question": "...",
-    "completion": "..."
-}
-```
-
-#### llama3.prompt_pairs
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "completion": "..."
-}
-```
-
-#### llama3.ultra
-
-```json
-{
-    "system": "...", // optional
-    "prompt": "...",
-    "completion": "..."
-}
-```
-
-#### user_defined.default
-
-For custom behaviors,
-
-```yaml
-rl: kto
-datasets:
-  - path: ...
-    split: train
-    type:
-      field_prompt: "prompt"
-      field_system: "system"
-      field_completion: "completion"
-      field_label: "label"
-      prompt_format: "{prompt}"
-      completion_format: "{completion}"
-```
-
-The input format is a simple JSON input with customizable fields based on the above config.
-
-```json
-{
-    "system": "...",  // optional
-    "prompt": "...",
-    "completion": "...",
-    "label": "..."
-}
-```
-
-### GRPO
-
-::: {.callout-tip}
-Check out our [GRPO cookbook](https://github.com/axolotl-ai-cloud/grpo_code).
-:::
-
-In the latest GRPO implementation, `vLLM` is used to significantly speedup trajectory generation during training. In this example, we're using 4 GPUs - 2 for training, and 2 for vLLM:
-
-::: {.callout-important}
-Make sure you've installed the correct version of vLLM by including it as an extra when installing axolotl, e.g. `pip install axolotl[vllm]`.
-:::
-
-```yaml
-base_model: Qwen/Qwen2.5-1.5B-Instruct
-
-vllm:
-    host: 0.0.0.0
-    port: 8000
-    tensor_parallel_size: 2
-    gpu_memory_utilization: 0.85
-    dtype: auto
-    # max_model_len: # you may find it useful to set the vLLM model context length if you know this beforehand
-
-rl: grpo
-trl:
-    use_vllm: true
-    vllm_server_host: 0.0.0.0
-    vllm_server_port: 8000
-    vllm_server_timeout: 300
-```
-
-```bash
-CUDA_VISIBLE_DEVICES=2,3 axolotl vllm-serve grpo.yaml
-```
-
-Your `vLLM` instance will now attempt to spin up, and it's time to kick off training utilizing our remaining two GPUs. In another terminal, execute:
-
-```bash
-CUDA_VISIBLE_DEVICES=0,1 axolotl train grpo.yaml --num-processes 2
-```
-
-::: {.callout-note}
-Due to TRL's implementation with vLLM, the vLLM instance must use the last N GPUs instead of the first N GPUs. This is why in the example above, we use `CUDA_VISIBLE_DEVICES=2,3` for the vLLM instance.
-:::
-
-#### Reward functions
-
-GRPO uses custom reward functions and transformations. Please have them ready locally.
-
-For example, to load OpenAI's GSM8K and use a random reward for completions:
-
-```python
-# rewards.py
-import random
-
-def rand_reward_func(completions, **kwargs) -> list[float]:
-    return [random.uniform(0, 1) for _ in completions]
-
-def oai_gsm8k_transform(cfg, *args, **kwargs):
-    def transform_fn(example, tokenizer=None):
-        label = example["answer"].split("####")[-1].strip().replace(",", "")
-        return {
-            "prompt": [{"role": "user", "content": example["question"]},],
-            "answer": label,
-        }
-    return transform_fn, {"remove_columns": ["question"]}
-```
-
-```yaml
-rl: grpo
-
-trl:
-    beta: 0.001
-    max_completion_length: 256
-    use_vllm: True
-    num_generations: 4
-    reward_funcs: ["rewards.rand_reward_func"]    # format: '{file_name}.{fn_name}'
-    reward_weights: [1.0]
-datasets:
-  - path: openai/gsm8k
-    name: main
-    type: rewards.oai_gsm8k_transform  # format: '{file_name}.{fn_name}'
-```
-
-To see other examples of custom reward functions, please see [TRL GRPO Docs](https://github.com/huggingface/trl/blob/main/docs/source/grpo_trainer.md#using-a-custom-reward-function).
-
-To see all configs, please see [TRLConfig](https://github.com/axolotl-ai-cloud/axolotl/blob/v0.9.2/src/axolotl/utils/schemas/trl.py).
-
-#### GRPO with DAPO/Dr. GRPO loss
-
-The DAPO paper and subsequently Dr. GRPO paper proposed an alternative loss function for GRPO to remediate the penalty in longer responses.
-
-```yaml
-trl:
-  loss_type: dr_grpo
-  # Normalizes loss based on max completion length (default: 256)
-  max_completion_length:
-```
-
-For more information, see [GRPO docs](https://huggingface.co/docs/trl/v0.17.0/en/grpo_trainer#loss-types).
-
-### SimPO
-
-SimPO uses [CPOTrainer](https://huggingface.co/docs/trl/main/en/cpo_trainer) but with alternative loss function.
-
-```yaml
-rl: simpo
-rl_beta: 0.1  # default in CPOTrainer
-cpo_alpha: 1.0  # default in CPOTrainer
-simpo_gamma: 0.5  # default in CPOTrainer
-```
-
-This method uses the same dataset format as [DPO](#dpo).
-
-### Using local dataset files
-
+#### Using local dataset files
 ```yaml
 datasets:
  - ds_type: json
@@ -619,9 +82,9 @@ datasets:
    type: chatml.intel
 ```

-### TRL auto-unwrapping for PEFT
+#### Trl autounwrap for peft

-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:
+Trl supports autounwrapping peft models, so that a ref model does not need to be additionally loaded, leading to less VRAM needed. This is on by default. To turn it off, pass the following config.

 ```yaml
 # load ref model when adapter training.
--- a/docs/scripts/generate_config_docs.py
+++ b/docs/scripts/generate_config_docs.py
@@ -1,752 +0,0 @@
-# type: ignore
-
-"""
-Quarto documentation generation from Pydantic models. Uses Pydantic model source code
-to automatically group fields, including inherited fields from parent classes.
-"""
-
-import ast
-import inspect
-import textwrap
-import types
-import typing
-from typing import Any, FrozenSet, Type, Union
-
-from pydantic import BaseModel
-
-from axolotl.utils.schemas.config import AxolotlInputConfig
-
-
-class QuartoGenerator:
-    """Generate Quarto documentation from Pydantic models."""
-
-    def __init__(self):
-        self._class_fields_cache = {}
-        self._inheritance_map_cache = {}
-        self._nested_models_cache = {}
-
-    def _get_direct_fields(self, cls: Type[BaseModel]) -> FrozenSet[str]:
-        """Get fields defined directly in a single class (not inherited)."""
-        if cls in self._class_fields_cache:
-            return self._class_fields_cache[cls]
-
-        fields = set()
-
-        # Get annotated fields
-        if hasattr(cls, "__annotations__"):
-            fields.update(cls.__annotations__.keys())
-
-        # Filter out private/special methods
-        fields = {f for f in fields if not f.startswith("_")}
-
-        result = frozenset(fields)
-        self._class_fields_cache[cls] = result
-        return result
-
-    def _is_pydantic_model(self, type_obj) -> bool:
-        """Check if a type is a Pydantic BaseModel."""
-        return inspect.isclass(type_obj) and issubclass(type_obj, BaseModel)
-
-    # pylint: disable=too-many-return-statements
-    def _extract_nested_type(self, field_type) -> Any:
-        """Extract the actual type from complex type annotations."""
-        # Handle Annotated types (Python 3.9+)
-        if hasattr(typing, "get_origin") and hasattr(typing, "get_args"):
-            origin = typing.get_origin(field_type)
-            args = typing.get_args(field_type)
-
-            if origin is not None:
-                # Handle Annotated[SomeType, ...] - extract the first argument
-                if hasattr(typing, "Annotated") and origin is typing.Annotated:
-                    if args:
-                        return self._extract_nested_type(
-                            args[0]
-                        )  # Recursively process the actual type
-
-                # Handle list[SomeType], List[SomeType], etc.
-                elif origin in (list, typing.List):
-                    if args:
-                        return self._extract_nested_type(
-                            args[0]
-                        )  # Extract element type
-
-                # Handle Union types (including | syntax)
-                elif origin is typing.Union:
-                    # Get non-None types from the Union
-                    non_none_types = [arg for arg in args if arg is not type(None)]
-                    if len(non_none_types) >= 1:
-                        # Prioritize Pydantic models over primitive types
-                        pydantic_models = [
-                            arg
-                            for arg in non_none_types
-                            if self._is_pydantic_model(arg)
-                        ]
-                        if pydantic_models:
-                            # Return the first Pydantic model found
-                            return self._extract_nested_type(pydantic_models[0])
-
-                        # No Pydantic models, return the first non-None type
-                        return self._extract_nested_type(non_none_types[0])
-
-        # Handle new Python 3.10+ union syntax (PeftConfig | None)
-        if hasattr(field_type, "__class__") and field_type.__class__ is types.UnionType:
-            # Get non-None types from the Union
-            non_none_types = [
-                arg for arg in field_type.__args__ if arg is not type(None)
-            ]
-            if len(non_none_types) >= 1:
-                # Prioritize Pydantic models over primitive types
-                pydantic_models = [
-                    arg for arg in non_none_types if self._is_pydantic_model(arg)
-                ]
-                if pydantic_models:
-                    return self._extract_nested_type(pydantic_models[0])
-                return self._extract_nested_type(non_none_types[0])
-
-        # Handle old typing.Union syntax (fallback)
-        if hasattr(field_type, "__origin__"):
-            if field_type.__origin__ is Union:
-                # Get non-None types from the Union
-                non_none_types = [
-                    arg for arg in field_type.__args__ if arg is not type(None)
-                ]
-                if len(non_none_types) >= 1:
-                    # Prioritize Pydantic models over primitive types
-                    pydantic_models = [
-                        arg for arg in non_none_types if self._is_pydantic_model(arg)
-                    ]
-                    if pydantic_models:
-                        return self._extract_nested_type(pydantic_models[0])
-                    return self._extract_nested_type(non_none_types[0])
-            # Handle other generic types like dict[str, Any], etc.
-            elif hasattr(field_type, "__args__"):
-                return field_type
-
-        return field_type
-
-    # pylint: disable=too-many-return-statements
-    def _extract_all_pydantic_models_from_type(
-        self, field_type
-    ) -> list[type[BaseModel]]:
-        """Extract all Pydantic models from a type annotation, including from Unions."""
-        models = []
-
-        if field_type is None:
-            return models
-
-        # Handle Annotated types
-        if hasattr(typing, "get_origin") and hasattr(typing, "get_args"):
-            origin = typing.get_origin(field_type)
-            args = typing.get_args(field_type)
-
-            if origin is not None:
-                # Handle Annotated[SomeType, ...] - extract from the first argument
-                if hasattr(typing, "Annotated") and origin is typing.Annotated:
-                    if args:
-                        models.extend(
-                            self._extract_all_pydantic_models_from_type(args[0])
-                        )
-                    return models
-
-                # Handle list[SomeType], List[SomeType], etc.
-                if origin in (list, typing.List):
-                    if args:
-                        models.extend(
-                            self._extract_all_pydantic_models_from_type(args[0])
-                        )
-                    return models
-
-                # Handle Union types
-                if origin is typing.Union:
-                    for arg in args:
-                        if arg is not type(None):  # Skip None type
-                            models.extend(
-                                self._extract_all_pydantic_models_from_type(arg)
-                            )
-                    return models
-
-        # Handle new Python 3.10+ union syntax
-        if hasattr(field_type, "__class__") and field_type.__class__ is types.UnionType:
-            for arg in field_type.__args__:
-                if arg is not type(None):  # Skip None type
-                    models.extend(self._extract_all_pydantic_models_from_type(arg))
-            return models
-
-        # Handle old typing.Union syntax (fallback)
-        if hasattr(field_type, "__origin__") and field_type.__origin__ is Union:
-            for arg in field_type.__args__:
-                if arg is not type(None):  # Skip None type
-                    models.extend(self._extract_all_pydantic_models_from_type(arg))
-            return models
-
-        # Check if this type itself is a Pydantic model
-        if self._is_pydantic_model(field_type):
-            models.append(field_type)
-
-        return models
-
-    def _get_nested_models(
-        self, model_class: type[BaseModel], visited=None
-    ) -> dict[str, type[BaseModel]]:
-        """Get all nested Pydantic models from a model class."""
-        if visited is None:
-            visited = set()
-
-        # Avoid infinite recursion
-        if model_class in visited:
-            return {}
-
-        if model_class in self._nested_models_cache:
-            return self._nested_models_cache[model_class]
-
-        visited.add(model_class)
-        nested_models = {}
-
-        # Check all fields in the model
-        for field_info in model_class.model_fields.values():
-            field_type = self._extract_nested_type(field_info.annotation)
-
-            if self._is_pydantic_model(field_type):
-                nested_models[field_type.__name__] = field_type
-                # Recursively get nested models from this nested model
-                deeper_nested = self._get_nested_models(field_type, visited.copy())
-                nested_models.update(deeper_nested)
-
-        self._nested_models_cache[model_class] = nested_models
-        return nested_models
-
-    def _build_inheritance_map(self, child_class: Type[BaseModel]):
-        """Build inheritance map for a class and all its parents."""
-        if child_class in self._inheritance_map_cache:
-            return self._inheritance_map_cache[child_class]
-
-        inheritance_map = {}
-
-        # Get MRO and filter out BaseModel and object
-        mro_classes = [
-            cls
-            for cls in child_class.__mro__
-            if cls not in (BaseModel, object) and hasattr(cls, "__annotations__")
-        ]
-
-        # Process each class in the MRO
-        for cls in mro_classes:
-            inheritance_map[cls] = self._get_direct_fields(cls)
-
-        self._inheritance_map_cache[child_class] = inheritance_map
-        return inheritance_map
-
-    def _wrap_comment(self, text: str, width: int = 88) -> list[str]:
-        """Wrap a comment to specified width, accounting for '# ' prefix."""
-        if not text.strip():
-            return ["#"]
-
-        # Account for "# " prefix (2 characters)
-        content_width = width - 2
-        wrapped_lines = textwrap.wrap(text, width=content_width)
-        return [f"# {line}" for line in wrapped_lines]
-
-    def _extract_type_from_source(
-        self, model_class: type[BaseModel], field_name: str
-    ) -> str:
-        """Extract the actual type annotation text from source code, checking inheritance chain."""
-        # Use inheritance map to check classes efficiently
-        inheritance_map = self._build_inheritance_map(model_class)
-
-        # Check classes in MRO order
-        for cls in model_class.__mro__:
-            if cls in inheritance_map and field_name in inheritance_map[cls]:
-                type_annotation = self._get_type_from_class_source(cls, field_name)
-                if type_annotation != "unknown":
-                    return type_annotation
-
-        return "unknown"
-
-    def _get_type_from_class_source(self, class_obj: type, field_name: str) -> str:
-        """Extract type annotation from a specific class's source code."""
-        try:
-            source = inspect.getsource(class_obj)
-            tree = ast.parse(source)
-        except (OSError, TypeError):
-            return "unknown"
-
-        # Find the class definition
-        for node in tree.body:
-            if isinstance(node, ast.ClassDef) and node.name == class_obj.__name__:
-                # Find the field assignment
-                for body_node in node.body:
-                    if isinstance(body_node, ast.AnnAssign) and isinstance(
-                        body_node.target, ast.Name
-                    ):
-                        if body_node.target.id == field_name and body_node.annotation:
-                            return ast.unparse(body_node.annotation)
-                break
-
-        return "unknown"
-
-    def _extract_field_groups_from_all_classes(
-        self, model_class: type[BaseModel]
-    ) -> list[dict]:
-        """Extract field groups from all classes in the inheritance hierarchy."""
-        all_groups = []
-        inheritance_map = self._build_inheritance_map(model_class)
-
-        # Get all Pydantic base classes in MRO order (most specific first)
-        # This puts AxolotlInputConfig fields first, then parent class fields
-        pydantic_classes = [
-            cls
-            for cls in model_class.__mro__
-            if cls in inheritance_map and inheritance_map[cls]
-        ]
-
-        # Extract groups from each class
-        for cls in pydantic_classes:
-            class_groups = self._extract_field_groups_from_source(cls)
-            for group in class_groups:
-                all_groups.append(group)
-
-        # If no groups found, create a default grouping by class
-        if not all_groups:
-            for cls in pydantic_classes:
-                fields_in_class = inheritance_map[cls]
-                if fields_in_class:
-                    all_groups.append(
-                        {
-                            "fields": list(fields_in_class),
-                        }
-                    )
-
-        return all_groups
-
-    # pylint: disable=too-many-return-statements
-    def _extract_field_groups_from_source(
-        self, model_class: type[BaseModel]
-    ) -> list[dict]:
-        """Extract field groups from source code based on blank lines and comments."""
-        try:
-            source = inspect.getsource(model_class)
-            tree = ast.parse(source)
-        except (OSError, TypeError):
-            # Fallback if we can't get source code
-            fields_in_class = self._get_direct_fields(model_class)
-            if fields_in_class:
-                return [
-                    {
-                        "fields": list(fields_in_class),
-                    }
-                ]
-            return []
-
-        groups = []
-        current_group_fields = []
-        current_group_comment = None
-
-        # Find the class definition
-        class_node = None
-        for node in ast.walk(tree):
-            if isinstance(node, ast.ClassDef) and node.name == model_class.__name__:
-                class_node = node
-                break
-
-        if not class_node:
-            fields_in_class = self._get_direct_fields(model_class)
-            if fields_in_class:
-                return [
-                    {
-                        "fields": list(fields_in_class),
-                    }
-                ]
-            return []
-
-        # Parse the source lines to detect groupings
-        source_lines = source.split("\n")
-
-        # Get fields that are actually defined in this specific class
-        fields_in_class = self._get_direct_fields(model_class)
-
-        # Find assignments that correspond to model fields for THIS class only
-        field_assignments = []
-        for node in class_node.body:
-            if isinstance(node, ast.AnnAssign) and isinstance(node.target, ast.Name):
-                field_name = node.target.id
-                if field_name in fields_in_class:
-                    field_assignments.append(
-                        {
-                            "name": field_name,
-                            "lineno": node.lineno,
-                            "end_lineno": getattr(node, "end_lineno", node.lineno),
-                        }
-                    )
-
-        if not field_assignments:
-            if fields_in_class:
-                return [
-                    {
-                        "fields": list(fields_in_class),
-                    }
-                ]
-            return []
-
-        # Sort by line number
-        field_assignments.sort(key=lambda x: x["lineno"])
-
-        # Group fields based on blank lines and comments
-        for i, field_info in enumerate(field_assignments):
-            field_name = field_info["name"]
-            current_line = field_info["lineno"]
-
-            # Check if this starts a new group (blank line before or significant gap)
-            is_new_group = False
-
-            if i == 0:
-                is_new_group = True
-            else:
-                prev_end_line = field_assignments[i - 1]["end_lineno"]
-
-                # Check for blank lines or comments between fields
-                lines_between = source_lines[prev_end_line : current_line - 1]
-                has_blank_line = any(line.strip() == "" for line in lines_between)
-                has_comment = any(
-                    line.strip().startswith("#") for line in lines_between
-                )
-
-                # Start new group if there's a blank line or comment, or significant gap
-                if has_blank_line or has_comment or (current_line - prev_end_line > 3):
-                    is_new_group = True
-
-            if is_new_group and current_group_fields:
-                # Save the previous group
-                groups.append(
-                    {
-                        "fields": current_group_fields.copy(),
-                        "description": current_group_comment,
-                    }
-                )
-                current_group_fields = []
-                current_group_comment = None
-
-            current_group_fields.append(field_name)
-
-        # Add the final group
-        if current_group_fields:
-            groups.append(
-                {
-                    "fields": current_group_fields,
-                    "description": current_group_comment,
-                }
-            )
-
-        return groups
-
-    def _generate_field_documentation(
-        self,
-        model_class: type[BaseModel],
-        field_name: str,
-        field_info: dict,
-        field_type_str: str,
-        is_required: bool,
-        indent_level: int = 0,
-        visited_models: set = None,
-    ) -> list[str]:
-        """Generate documentation for a single field, expanding nested models inline."""
-        if visited_models is None:
-            visited_models = set()
-
-        lines = []
-        indent = "  " * indent_level
-
-        # Get the actual field type for nested model detection
-        if field_name in model_class.model_fields:
-            pydantic_field_info = model_class.model_fields[field_name]
-            actual_field_type = pydantic_field_info.annotation
-        else:
-            actual_field_type = None
-
-        # Add description comment if available
-        description = field_info.get("description", "")
-        if description:
-            wrapped_lines = self._wrap_comment(description, width=88 - len(indent))
-            for line in wrapped_lines:
-                lines.append(f"{indent}{line}")
-
-        # Extract nested Pydantic models from the type annotation
-        nested_models = self._extract_all_pydantic_models_from_type(actual_field_type)
-
-        # Filter out already visited models to prevent infinite recursion
-        expandable_models = [
-            model for model in nested_models if model not in visited_models
-        ]
-
-        if expandable_models:
-            # This field contains Pydantic models that can be expanded
-
-            # Show the field with its full type annotation
-            field_line = f"{indent}{field_name}: {field_type_str}"
-            if field_info.get("default") is not None:
-                field_line += f" = {field_info['default']}"
-            if is_required:
-                field_line += " (required)"
-            lines.append(field_line)
-
-            # Add to visited to prevent infinite recursion
-            new_visited = visited_models.copy()
-            new_visited.update(expandable_models)
-
-            # Expand each nested Pydantic model
-            for i, nested_model in enumerate(expandable_models):
-                if i > 0:
-                    lines.append("\n")
-                lines.append(f"{indent}  # For {nested_model.__name__}:")
-
-                # Get nested model schema
-                try:
-                    nested_schema = nested_model.model_json_schema()
-                    nested_properties = nested_schema.get("properties", {})
-                    nested_required = nested_schema.get("required", [])
-                except Exception:  # pylint: disable=broad-exception-caught
-                    # Fallback: use model fields directly
-                    nested_properties = {}
-                    nested_required = []
-                    for (
-                        nested_field_name,
-                        nested_field_info,
-                    ) in nested_model.model_fields.items():
-                        nested_description = ""
-                        if (
-                            hasattr(nested_field_info, "json_schema_extra")
-                            and nested_field_info.json_schema_extra
-                        ):
-                            nested_description = (
-                                nested_field_info.json_schema_extra.get(
-                                    "description", ""
-                                )
-                            )
-                        elif (
-                            hasattr(nested_field_info, "description")
-                            and nested_field_info.description
-                        ):
-                            nested_description = nested_field_info.description
-
-                        nested_default_val = None
-                        if (
-                            hasattr(nested_field_info, "default")
-                            and nested_field_info.default is not None
-                        ):
-                            if str(nested_field_info.default) != "PydanticUndefined":
-                                nested_default_val = nested_field_info.default
-
-                        nested_properties[nested_field_name] = {
-                            "type": "unknown",
-                            "description": nested_description,
-                            "default": nested_default_val,
-                        }
-
-                        if nested_field_info.is_required():
-                            nested_required.append(nested_field_name)
-
-                # Get field groups for the nested model
-                nested_field_groups = self._extract_field_groups_from_all_classes(
-                    nested_model
-                )
-
-                # Generate nested fields with increased indentation
-                for i, group in enumerate(nested_field_groups):
-                    if not group["fields"]:
-                        continue
-
-                    # Add blank line between groups (except before first group)
-                    if i > 0:
-                        lines.append("")
-
-                    # Process nested fields
-                    for nested_field_name in group["fields"]:
-                        if nested_field_name not in nested_properties:
-                            continue
-
-                        nested_field_info = nested_properties[nested_field_name]
-                        nested_field_type = self._extract_type_from_source(
-                            nested_model, nested_field_name
-                        )
-                        nested_is_required = nested_field_name in nested_required
-
-                        # Recursively generate documentation for nested field
-                        nested_lines = self._generate_field_documentation(
-                            nested_model,
-                            nested_field_name,
-                            nested_field_info,
-                            nested_field_type,
-                            nested_is_required,
-                            indent_level + 1,
-                            new_visited,
-                        )
-                        lines.extend(nested_lines)
-        else:
-            # Regular field (no expandable nested models)
-            field_line = f"{indent}{field_name}: {field_type_str}"
-            if field_info.get("default") is not None:
-                field_line += f" = {field_info['default']}"
-            if is_required:
-                field_line += " (required)"
-            lines.append(field_line)
-
-        return lines
-
-    def generate_qmd(
-        self,
-        model_class: type[BaseModel],
-        title: str | None = None,
-        expand_nested: bool = True,
-    ) -> str:
-        """Auto-generate config reference documentation including inherited fields."""
-
-        if title is None:
-            title = f"{model_class.__name__} Reference"
-
-        # Try to get JSON schema, with fallback for serialization issues
-        try:
-            schema = model_class.model_json_schema()
-            properties = schema.get("properties", {})
-            required = schema.get("required", [])
-        except Exception as e:  # pylint: disable=broad-exception-caught
-            print(
-                f"Warning: Could not generate JSON schema ({e}). Using model fields instead."
-            )
-            # Fallback: use model fields directly
-            properties = {}
-            required = []
-            for field_name, field_info in model_class.model_fields.items():
-                # Extract description from json_schema_extra or field info
-                description = ""
-                if (
-                    hasattr(field_info, "json_schema_extra")
-                    and field_info.json_schema_extra
-                ):
-                    description = field_info.json_schema_extra.get("description", "")
-                elif hasattr(field_info, "description") and field_info.description:
-                    description = field_info.description
-
-                # Get default value
-                default_val = None
-                if hasattr(field_info, "default") and field_info.default is not None:
-                    # Handle special Pydantic default markers
-                    if str(field_info.default) != "PydanticUndefined":
-                        default_val = field_info.default
-
-                properties[field_name] = {
-                    "type": "unknown",
-                    "description": description,
-                    "default": default_val,
-                }
-
-                if field_info.is_required():
-                    required.append(field_name)
-
-        # Extract field groups from all classes in inheritance hierarchy
-        field_groups = self._extract_field_groups_from_all_classes(model_class)
-
-        # Start building QMD content
-        qmd_lines = [
-            "---",
-            f"title: {title}",
-            "description: A complete list of all configuration options.",
-            "---",
-            "",
-        ]
-
-        # Generate one big code block with all fields (inline nested expansion)
-        qmd_lines.append("```yaml")
-
-        for i, group in enumerate(field_groups):
-            if not group["fields"]:
-                continue
-
-            # Add blank line between groups (except before first group)
-            if i > 0:
-                qmd_lines.append("")
-
-            # Process fields in the order they appear in source
-            for field_name in group["fields"]:
-                if field_name not in properties:
-                    continue
-
-                field_info = properties[field_name]
-                field_type = self._extract_type_from_source(model_class, field_name)
-                is_required = field_name in required
-
-                if expand_nested:
-                    # Check if this field has nested models
-                    if field_name in model_class.model_fields:
-                        pydantic_field_info = model_class.model_fields[field_name]
-                        nested_models = self._extract_all_pydantic_models_from_type(
-                            pydantic_field_info.annotation
-                        )
-                        has_nested = bool(nested_models)
-                    else:
-                        has_nested = False
-
-                    # Add blank line before nested config
-                    if has_nested:
-                        qmd_lines.append("")
-
-                    # Use the new inline generation method
-                    field_lines = self._generate_field_documentation(
-                        model_class,
-                        field_name,
-                        field_info,
-                        field_type,
-                        is_required,
-                        indent_level=0,
-                        visited_models=set(),
-                    )
-                    qmd_lines.extend(field_lines)
-
-                    # Add blank line after nested config
-                    if has_nested:
-                        qmd_lines.append("")
-                else:
-                    # Original simple approach
-                    description = field_info.get("description", "")
-                    default = field_info.get("default")
-
-                    # Add wrapped comment for description
-                    if description:
-                        wrapped_lines = self._wrap_comment(description)
-                        qmd_lines.extend(wrapped_lines)
-
-                    line = f"{field_name}: {field_type}"
-                    if default is not None:
-                        line += f" = {default}"
-                    if is_required:
-                        line += " (required)"
-                    qmd_lines.append(line)
-
-        qmd_lines.append("```")
-
-        # Join all lines and clean up any double newlines
-        content = "\n".join(qmd_lines)
-
-        # Replace multiple consecutive newlines with just two newlines (one blank line)
-        import re
-
-        content = re.sub(r"\n{3,}", "\n\n", content)
-
-        # Ensure single newline at the very end
-        content = content.rstrip("\n") + "\n"
-
-        return content
-
-
-def main():
-    generator = QuartoGenerator()
-
-    print("Generating config reference content...")
-    qmd_content = generator.generate_qmd(AxolotlInputConfig, "Config Reference", True)
-
-    print("Writing to file...")
-    with open("docs/config-reference.qmd", "w", encoding="utf-8") as f:
-        f.write(qmd_content)
-    print("Done!")
-
-
-if __name__ == "__main__":
-    main()
--- a/docs/sequence_parallelism.qmd
+++ b/docs/sequence_parallelism.qmd
@@ -1,100 +0,0 @@
---
-title: Sequence Parallelism
-description: Train with long sequences split across multiple GPUs.
---
-
-Sequence parallelism is a technique that splits sequences across multiple GPUs,
-allowing you to train with very long sequences that wouldn't fit on a single GPU. Each
-GPU processes a different portion of the sequence, and the results are aggregated
-through a ring communication pattern.
-
-## When to Use Sequence Parallelism
-
-Use sequence parallelism when:
-
- You need to train with sequence lengths that don't fit into a single GPU's memory
- You have multiple GPUs available
- You're experiencing OOM (Out Of Memory) errors with long sequences
-
-## Configuration
-
-To enable sequence parallelism, add the following to your configuration file:
-
-```yaml
-# Set to a divisor (> 1) of the number of GPUs available
-context_parallel_size: 4  # Split sequences across 4 GPUs
-# Optional; strides across the key dimension. Larger values use more memory but should make training faster.
-heads_k_stride: 1
-# Optional; one of "varlen_llama3" or "batch_ring". Defaults to
-# "varlen_llama3" when `sample_packing: true`, and "batch_ring" otherwise.
-ring_attn_func:
-```
-
-The `context_parallel_size` should be a divisor of the total number of GPUs. For example:
-
- With 8 GPUs, valid values would be 2, 4, or 8
- With 4 GPUs, valid values would be 2 or 4
-
-## Implementation Details
-
-When sequence parallelism is enabled:
-
-1. Each sequence is divided into equal chunks across the GPUs in a sequence parallel group
-2. The data collator handles the chunking of input_ids, attention_mask, labels, and position_ids
-3. Position IDs are adjusted to maintain proper relative positions
-4. The trainer uses special ring communication patterns for attention operations
-
-## Requirements
-
-To use sequence parallelism, you need:
-
- Multiple GPUs (at least 2)
- The `ring-flash-attn` package. Install with:
-  - `pip install axolotl[ring-flash-attn]` (preferred)
-  - `pip install ring-flash-attn>=0.1.4`
-
-## Limitations
-
- Flash attention must be enabled for this to work (`flash_attention: true` in config YAML)
- May have a small performance overhead due to communication between GPUs
-
-## Example
-
-```yaml
-base_model: meta-llama/Llama-3-8B-Instruct
-sequence_len: 8192
-
-...
-
-context_parallel_size: 4  # Split each sequence into 4 parts, one per GPU
-# Optional; strides across the key dimension. Larger values use more memory but should make training faster.
-heads_k_stride: 1
-# Optional; one of "varlen_llama3" or "batch_ring". Defaults to
-# "varlen_llama3" when `sample_packing: true`, and "batch_ring" otherwise.
-ring_attn_func:
-
-...
-```
-
-This will train the Llama 3 8B model with 8K context length, with each sequence split
-into 2 subsequences of length 4096 across 2 GPUs.
-
-## Sample Packing with Sequence Parallelism
-
-Sequence parallelism is compatible with Axolotl's sample packing functionality. When using both features together:
-
-1. Samples are first packed together
-2. The packed sequences are then divided across GPUs in the sequence parallel group
-3. Position IDs are automatically adjusted to maintain proper relative positions
-
-## Effect on Batch Size
-
-When using sequence parallelism, your effective global batch size is **divided** by the `context_parallel_size`. This happens because:
-
- Each group of `context_parallel_size` GPUs works on the same batch (just different parts of each sequence)
- The number of batches processed per step decreases
-
-For example:
- With 8 GPUs and no sequence parallelism: 8 different batches processed per step
- With 8 GPUs and `context_parallel_size=4`: Only 2 different batches processed per step (each split across 4 GPUs)
- If your per-GPU `micro_batch_size` is 2, the global batch size decreases from 16 to 4
--- a/docs/torchao.qmd
+++ b/docs/torchao.qmd
@@ -3,12 +3,6 @@ title: "PyTorch ao"
 description: "Custom data types and layouts for training and inference"
 ---

-To use experimental optimizers (`AdamWFp8`, `AdamW4bit`, `AdamW8bit`) from Pytorch Ao, please install the package as shown below.
-
-::: {.callout-tip}
-Some experimental optimizers are already present in regular Pytorch, so please re-check if you actually need this package!
-:::
-
 ### Installation

 Stable Release from the PyTorch index
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Wing Lian	8c4f89745a	fix softmax class check	2025-01-15 23:23:13 -05:00
Wing Lian	36b71f34d7	register rala	2025-01-15 23:21:22 -05:00
Wing Lian	d28fee7609	use autoconfig w rala	2025-01-15 23:14:47 -05:00
Wing Lian	c196776996	option to not concatenate during pretraining	2025-01-15 22:45:02 -05:00
Wing Lian	79ae776102	fixup logging layer	2025-01-15 21:36:14 -05:00
Wing Lian	145664d82c	more fixups	2025-01-15 21:27:12 -05:00
Dan Saunders	28694219a5	inline comment change	2025-01-14 16:59:43 +00:00
Dan Saunders	fd8ad6fcbf	fixing negative component mixing	2025-01-13 19:21:55 +00:00
Dan Saunders	661d71a14b	adding diff attn negative component warmup (in progress)	2025-01-10 21:57:31 +00:00
Dan Saunders	6dd47edcb8	fire CLI fixes	2025-01-10 18:24:16 +00:00
Dan Saunders	7aca08ff60	adding guard statements	2025-01-10 16:39:21 +00:00
Dan Saunders	4f804f6d88	adding diff attn callback, adding documentation	2025-01-10 16:28:51 +00:00
Dan Saunders	443327c585	CLI build_command bugfix	2025-01-10 16:28:51 +00:00
Dan Saunders	70c4e6fbe6	updates and cleanup	2025-01-10 16:28:51 +00:00
Dan Saunders	2a7f139ad2	pre-commit fix	2025-01-10 16:28:51 +00:00
Dan Saunders	332ce0ae85	fixes and cleanup	2025-01-10 16:28:51 +00:00
Dan Saunders	e5fa842ff8	update	2025-01-10 16:28:51 +00:00
Dan Saunders	78e0ec0aa5	changes	2025-01-10 16:28:51 +00:00
Dan Saunders	3bc568eb27	adding registration function	2025-01-10 16:28:51 +00:00
Dan Saunders	eb6611d55f	progress on modeling code	2025-01-10 16:28:51 +00:00
Dan Saunders	4ff3328e66	updated custom modeling code	2025-01-10 16:28:51 +00:00
Dan Saunders	a3fd5074a9	fix duplicate-code warnings	2025-01-10 16:28:51 +00:00
Dan Saunders	5b90da0be3	added modeling code; cleanup + refactor	2025-01-10 16:28:51 +00:00
Dan Saunders	fcbfa86373	refactor and fixing test isolation issues	2025-01-10 16:28:51 +00:00
Dan Saunders	0d56582090	adding yaml dumper preserving input config format	2025-01-10 16:28:51 +00:00
Dan Saunders	390cb5742e	removing extra pytest xdist args	2025-01-10 16:28:51 +00:00
Dan Saunders	1d935f65c3	moving tests around for flash_attn install	2025-01-10 16:28:51 +00:00
Dan Saunders	66176b3e07	adding split_heads argument for retaining original (Q, K) dimensionanlity	2025-01-10 16:28:51 +00:00
Dan Saunders	505321ac95	isolating problematic test	2025-01-10 16:28:51 +00:00
Dan Saunders	0b382c88da	fixes post-rebase	2025-01-10 16:28:51 +00:00
Dan Saunders	ea07a7086e	plugin implementation	2025-01-10 16:28:51 +00:00
Dan Saunders	d22e1136bc	convert-differential-transformer test coverage	2025-01-10 16:28:51 +00:00
Dan Saunders	63b8e42c6b	duplicate code ignore	2025-01-10 16:28:51 +00:00
Dan Saunders	bda1eed59e	differential flash attention 2; cleanup	2025-01-10 16:28:51 +00:00
Dan Saunders	41ebd93158	moving monkeypatch	2025-01-10 16:28:51 +00:00
Dan Saunders	4c050ce807	pre-commit fix	2025-01-10 16:28:51 +00:00
Dan Saunders	6665acf63d	fix model save / load logic	2025-01-10 16:28:51 +00:00
Dan Saunders	2f9fa4c465	various improvemnents	2025-01-10 16:28:51 +00:00
Dan Saunders	849bc94112	various improvemnents	2025-01-10 16:28:51 +00:00
Dan Saunders	e484ec778d	training fixes, patching, minor cleanup	2025-01-10 16:28:51 +00:00
Dan Saunders	df1504ae14	adding CLI command for convert-diff-transformer	2025-01-10 16:28:51 +00:00
Dan Saunders	7be0d7496c	Adding script for doing conversion; fixes and updates	2025-01-10 16:28:51 +00:00
Dan Saunders	13cdffa91f	initial diff attn layer / model conversion implementation (support for llama arch)	2025-01-10 16:28:51 +00:00
Dan Saunders	7a4b296f60	Basic evaluate CLI command / codepath (#2188 ) * basic evaluate CLI command / codepath * tests for evaluate CLI command * fixes and cleanup * review comments; slightly DRYing up things --------- Co-authored-by: Dan Saunders <danjsaund@gmail.com>	2025-01-10 16:28:51 +00:00