Skip to content

Add warning about --dtype arg in reference_hf.py. #10

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Add warning about --dtype arg in reference_hf.py. #10

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion references/supported_models.html
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -581,6 +581,10 @@ <h4>Interactive debugging<a class="headerlink" href="#interactive-debugging" tit
<li><p>Get the reference output by <code class="docutils literal notranslate"><span class="pre">python3</span> <span class="pre">scripts/playground/reference_hf.py</span> <span class="pre">--model-path</span> <span class="pre">[new</span> <span class="pre">model]</span> <span class="pre">--model-type</span> <span class="pre">{text,vlm}</span></code></p></li>
<li><p>Get the SGLang output by <code class="docutils literal notranslate"><span class="pre">python3</span> <span class="pre">-m</span> <span class="pre">sglang.bench_one_batch</span> <span class="pre">--correct</span> <span class="pre">--model</span> <span class="pre">[new</span> <span class="pre">model]</span></code></p></li>
</ul>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>When comparing outputs between reference_hf.py and SGLang, ensure that parameters such as <code class="docutils literal notranslate"><span class="pre">--dtype</span></code>, temperature, top_p, and other sampling parameters are identical across both scripts. Even small differences in these parameters can lead to variations in output that might be mistakenly attributed to implementation issues.</p>
</div>
</section>
<section id="add-the-model-to-the-test-suite">
<h4>Add the model to the test suite<a class="headerlink" href="#add-the-model-to-the-test-suite" title="Link to this heading">#</a></h4>
Expand Down Expand Up @@ -752,4 +756,4 @@ <h3>Registering an external model implementation<a class="headerlink" href="#reg
<footer class="bd-footer">
</footer>
</body>
</html>
</html>