tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Built site for gh-pages

Browse Source
This commit is contained in:
Quarto GHA Workflow Runner
2026-01-27 22:15:35 +00:00
parent 4c48b9b508
commit 4934c2f06a
9 changed files with 812 additions and 842 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
docs/api/cli.merge_sharded_fsdp_weights.html
View File

@@ -852,12 +852,10 @@ gtag('config', 'G-9KYCVJBNMQ', { 'anonymize_ip': true});
<div class="code-copy-outer-scaffold"><div class="sourceCode" id="cb3"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a>cli.merge_sharded_fsdp_weights.merge_fsdp_weights(</span>
<span id="cb3-2"><a href="#cb3-2" aria-hidden="true" tabindex="-1"></a> checkpoint_dir,</span>
<span id="cb3-3"><a href="#cb3-3" aria-hidden="true" tabindex="-1"></a> output_path,</span>
<span id="cb3-4"><a href="#cb3-4" aria-hidden="true" tabindex="-1"></a> safe_serialization<span class="op">=</span><span class="va">False</span>,</span>
<span id="cb3-5"><a href="#cb3-5" aria-hidden="true" tabindex="-1"></a> remove_checkpoint_dir<span class="op">=</span><span class="va">False</span>,</span>
<span id="cb3-6"><a href="#cb3-6" aria-hidden="true" tabindex="-1"></a>)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<span id="cb3-4"><a href="#cb3-4" aria-hidden="true" tabindex="-1"></a> remove_checkpoint_dir<span class="op">=</span><span class="va">False</span>,</span>
<span id="cb3-5"><a href="#cb3-5" aria-hidden="true" tabindex="-1"></a>)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<p>Merge the weights from sharded FSDP model checkpoints into a single combined checkpoint. Should be used if
<code>SHARDED_STATE_DICT</code> was used for the model. Weights will be saved to <code>{output_path}/model.safetensors</code> if
<code>safe_serialization</code> else <code>pytorch_model.bin</code>.</p>
<code>SHARDED_STATE_DICT</code> was used for the model. Weights will be saved to <code>{output_path}/model.safetensors</code>.</p>
<p>Note: this is a CPU-bound process.</p>
<section id="parameters-1" class="level4 doc-section doc-section-parameters">
<h4 class="doc-section doc-section-parameters anchored" data-anchor-id="parameters-1">Parameters</h4>
@@ -890,12 +888,6 @@ gtag('config', 'G-9KYCVJBNMQ', { 'anonymize_ip': true});
<td><em>required</em></td>
</tr>
<tr class="odd">
<td>safe_serialization</td>
<td><code>bool</code>, <em>optional</em>, defaults to <code>True</code></td>
<td>Whether to save the merged weights with safetensors (recommended).</td>
<td><code>False</code></td>
</tr>
<tr class="even">
<td>remove_checkpoint_dir</td>
<td><code>bool</code>, <em>optional</em>, defaults to <code>False</code></td>
<td>Whether to remove the checkpoint directory after merging.</td>