[Docs] Add Entrypoints section to User Guides#2518
Conversation
Adds detailed documentation for both PTQ entrypoints: - guides/entrypoints/index.md: decision table for oneshot vs model_free_ptq - guides/entrypoints/oneshot.md: full lifecycle, all arguments in tables, calibration pipelines, and examples (FP8, GPTQ W4A16, MoE) - guides/entrypoints/model-free-ptq.md: when to use, how it works internally, standard vs NVFP4 microscale flow (with reindexing), ignore patterns, and supported schemes Updates .nav.yml to nest the three pages under Entrypoints in User Guides. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
|
👋 Hi! Thank you for contributing to llm-compressor. Please add the ready label when the PR is ready for review. Note: This is required to complete the testing suite, please only add the label once the PR is code complete and local testing has been performed. |
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request significantly enhances the documentation for LLM Compressor's post-training quantization (PTQ) entrypoints. It introduces a dedicated 'Entrypoints' section within the User Guides, offering comprehensive details on both Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces new documentation for two post-training quantization (PTQ) entrypoints: oneshot and model_free_ptq. The oneshot guide details its use for schemes requiring calibration data and Hugging Face model definitions, while model_free_ptq covers data-free quantization for models without such definitions. The navigation file docs/.nav.yml has been updated to include these new guides. Feedback suggests improving the model_free_ptq documentation by updating the default value for the ignore argument to [] for better readability and formatting a note as a !!! note block for consistency. Additionally, a redundant example in the oneshot documentation should be removed.
Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
|
Tracked in JIRA: https://issues.redhat.com/browse/INFERENG-5605 JIRA Details:
|
brian-dellabetta
left a comment
brian-dellabetta
left a comment
There was a problem hiding this comment.
One nit, otherwise LGTM
|
@mergify queue |
Merge Queue Status
Required conditions to enter a queue
|
|
Clarified the description of the 'oneshot' usage regarding INT8 quantization. Signed-off-by: Dipika Sikka <dipikasikka1@gmail.com>
3 participants
Summary
Adds a new Entrypoints section under User Guides with detailed documentation for both PTQ entrypoints:
guides/entrypoints/index.md) — decision table comparingoneshotvsmodel_free_ptqto help users choose the right entrypointguides/entrypoints/oneshot.md) — full lifecycle (preprocessing, calibration, postprocessing), all arguments organized by category (model, recipe, dataset, pipeline, misc), calibration pipeline descriptions, and examples for FP8 data-free, GPTQ W4A16, and Llama4 MoE NVFP4 with a proper ignore listguides/entrypoints/model-free-ptq.md) — when to use (data-free schemes, no transformers definition, oneshot fallback), how it works internally (file-by-file safetensors processing), standard flow vs NVFP4 microscale flow (withreindex_fused_weights), ignore patterns, and supported schemes tableAlso updates
.nav.ymlto nest the three pages underEntrypointsin User Guides.🤖 Generated with Claude Code