AWQ minor performance improvements to smoothing

[brian-dellabetta]

SUMMARY:
I wanted to investigate how the runtime for apply_smoothing in AWQ could be decreased with torch.compile and removing python for loops. These changes reduce the runtime of apply_smoothing by about 8-10%, by vectorizing the loss computation.

TEST PLAN:
Confirmed wikitext PPL scores are virtually the same when running AWQ on "meta-llama/Llama-3.2-3B-Instruct" and "Qwen/Qwen2.5-7B-Instruct"

[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 @brian-dellabetta, 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 focuses on enhancing the performance of the AWQ smoothing process. It achieves this by refactoring how intermediate outputs are handled to enable vectorized loss computation and by leveraging torch.compile for key operations, resulting in a reported 8-10% reduction in apply_smoothing runtime.

Highlights

• Performance Improvement: Vectorized the loss computation within apply_smoothing by concatenating intermediate outputs into a single tensor, eliminating Python for loops and reducing runtime by 8-10%.
• Code Optimization: Applied @torch.compile() to the new _compute_loss function and the existing _pseudo_quantize_tensor function for potential runtime acceleration.
• Refactoring: Replaced the _run_samples method with a new _get_flattened_output function that returns a single, flattened tensor, streamlining data handling for loss calculation.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 is currently in preview and 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 to provide feedback.

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 successfully implements performance improvements for the AWQ smoothing process by vectorizing the loss computation and utilizing torch.compile. The refactoring to handle concatenated and flattened outputs for loss calculation is a good approach for efficiency. The changes appear correct and align with the stated goals. Minor improvements to docstrings could further enhance code clarity.