[bugfix] fix infer pt dp

[Jintao-Huang]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @Jintao-Huang, 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 introduces a bug fix primarily focused on enhancing the stability and correctness of the inference pipeline, particularly in distributed data parallel environments. The changes ensure that the system can robustly handle datasets of varying sizes during parallel processing and correctly aggregate and log inference outputs to JSONL files, leading to more reliable and consistent results.

Highlights

• Data Parallelism Robustness: Improved the inference process to gracefully handle scenarios where the validation dataset is smaller than the configured data parallel size, preventing potential errors during sharding.
• JSONL Output Management: Refactored the logic for writing inference results to a JSONL file, moving the append operation from the _batch_infer method to the infer_dataset method. This ensures proper aggregation of results across different inference batches.

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 fixes a bug in data parallel inference, specifically for cases where the dataset size is smaller than the number of parallel workers. The changes correctly handle data sharding by adjusting the data parallel size and allowing ranks with no data to exit early. Additionally, the code is refactored to move the logic for writing results to a JSONL file from the _batch_infer method to its caller, infer_dataset, which improves separation of concerns. The changes are correct and improve both robustness and code structure.

[gemini-code-assist]

For better readability and to make the early exit more explicit, you could check for the rank and exit before re-assigning data_parallel_size. This doesn't change the logic but can make it easier to follow.

Suggested change

	if len(val_dataset) < data_parallel_size:
	data_parallel_size = len(val_dataset)
	if rank >= len(val_dataset):
	return []
	if len(val_dataset) < data_parallel_size:
	if rank >= len(val_dataset):
	return []
	data_parallel_size = len(val_dataset)