Update lm-eval set-up to address regression

[dsikka]

SUMMARY:

• Seems like setting the collator from tuncation to default + shuffling addresses the regression we're seeing in lm-eval
• Given the recovery values you see in these tests were determined using these settings, I think they should be how we evaluate our lm-eval tests for the time being

[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 @dsikka, 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 resolves a regression encountered in lm-eval by adjusting the data processing configuration. The core change involves setting the data_collator to DefaultDataCollator and enabling the shuffling of calibration samples within the oneshot_kwargs, which collectively addresses the observed issue and ensures correct evaluation behavior.

Highlights

• Regression Fix: Addresses a regression observed in lm-eval by modifying the data collator setup.
• Data Collator Configuration: The data_collator in oneshot_kwargs is now explicitly set to DefaultDataCollator().
• Calibration Sample Shuffling: Enabled shuffling of calibration samples by setting shuffle_calibration_samples to True within oneshot_kwargs.

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 aims to fix a regression in lm-eval by changing the data collator used during the oneshot process. The changes involve updating the e2e testing utility to use DefaultDataCollator and enable shuffling of calibration samples. My review identifies a significant issue where the data_collator is unconditionally overwritten, which would break tests for datasets that use a custom collator. I've provided a suggestion to fix this by only setting the default collator if one isn't already present, ensuring that existing specialized test configurations for datasets like 'flickr30k' and 'calibration' continue to function correctly.

[HDCharles]

I'm unsure if we should make this change, if the new approach is lower variance and we're confident it's not an actual regression, should we just update the thresholds?

[dsikka]

I'm unsure if we should make this change, if the new approach is lower variance and we're confident it's not an actual regression, should we just update the thresholds?

Replied on slack

[HDCharles]

ok, we can make this change for now and then decide what we want long term later

[brian-dellabetta]

Nice catch! One breakpoint to remove