[AWQ][Qwen3 VL] Add qwen3-vl-30b-a3b-Instruct-example

[JartX]

You can follow the thread here:

#1939 (comment)

After confirming with @dsikka that the issue when quantizing the model was a datasheet problem, and that the flickr30k dataset / processing you were trying also results in the same failure (which specifically comes from transformers, suggesting there may be a mismatch in the flickr30k processing and what is needed by the Qwen3 VL Model), I finished a functional model without a datafree pipeline, but a sequential one, as you requested. I proceed to publish, as promised, both the quantization script and the already uploaded model here: https://huggingface.co/jart25/Qwen3-VL-30B-A3B-Instruct-AWQ-8bit

[github-actions]

👋 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.

[gemini-code-assist]

Summary of Changes

Hello @JartX, 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 adds a comprehensive example for quantizing the Qwen3-VL-30B-A3B-Instruct model using the AWQ method. The new script provides a working sequential pipeline, including model loading, data preprocessing for calibration, detailed AWQ recipe configuration with smoothing and balancing, and a demonstration of model generation and saving. This example aims to resolve previous quantization issues and offers a ready-to-use solution for users.

Highlights

• New Example for Qwen3-VL-30B-A3B-Instruct: Introduces a Python script (qwen3-vl-30b-a3b-Instruct-example.py) demonstrating AWQ quantization for the Qwen/Qwen3-VL-30B-A3B-Instruct model.
• Sequential Quantization Pipeline: Implements a functional quantization pipeline that is sequential rather than datafree, addressing previous challenges encountered during model quantization.
• Detailed AWQ Configuration: Includes a specific AWQ recipe with ignore patterns for certain layers and mappings for smoothing and balancing, utilizing 8-bit symmetric integer weights with a group size of 32.
• Pre-published Model and Script: The author has already published the quantized model and the associated script to Hugging Face, making it readily available for use.

Using Gemini Code Assist

The 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 /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

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 .gemini/ folder in the base of the repository. Detailed instructions can be found here.

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.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request adds a new example script for quantizing the Qwen3-VL-30B-A3B-Instruct model using AWQ. The script is a good addition, but I've found a few issues in the AWQ configuration that would prevent it from working correctly. Most critically, the ignore list is too restrictive and disables smoothing altogether. Additionally, the mappings are not correct for a Mixture-of-Experts model. I've also included some suggestions to improve code style and efficiency. Please review the detailed comments.