[Args] Shuffle data samples by default

[kylesayrs]

Purpose

• Better accuracy recovery is seen when shuffling data samples
• Assume that for batch_size=1, data_collator_with_truncation is equivalent to default_data_collator

[gemini-code-assist]

Summary of Changes

Hello @kylesayrs, 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 refines the model calibration process by changing the default behavior to shuffle calibration samples, which can improve the robustness of calibration. Additionally, it introduces a suite of new parameters to the oneshot entrypoint, offering enhanced control and flexibility over various calibration pipeline configurations, including options for tracing and sequential processing.

Highlights

• Default shuffle_calibration_samples changed to True: The default value for the shuffle_calibration_samples argument has been updated from False to True in both DatasetArguments and the oneshot entrypoint. This ensures that calibration samples are shuffled by default, promoting better data randomness during calibration.
• New calibration pipeline arguments introduced: Several new arguments related to calibration pipelines have been added to the oneshot entrypoint. These include pipeline for selecting the calibration strategy, tracing_ignore for specifying functions to skip during tracing, sequential_targets for defining layers in a sequential pipeline, and sequential_offload_device for managing intermediate activation offloading.

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 changes the default value for shuffle_calibration_samples to True and adds several new arguments to the oneshot entrypoint to support different calibration pipelines. The changes are mostly straightforward, but I've identified a significant issue with the use of a mutable list as a default argument for the new tracing_ignore parameter. This is a common pitfall in Python and should be addressed to avoid potential bugs. My review includes a detailed explanation and a recommendation for fixing it.

[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.

[brian-dellabetta]

one nit