[bugfix] Fix multimodal lazy_tokenize false

[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 addresses a bug related to multimodal lazy tokenization by refining how LazyLLMDataset objects are processed within dataset statistics. It also includes a critical patch for the Megatron framework's dataset building mechanism to ensure seamless integration and correct data handling across different training stages. Additionally, the changes improve diagnostic capabilities by adding a log for evaluation iteration settings.

Highlights

• Multimodal Lazy Tokenization Fix: The _stat_dataset method in sft.py has been updated to correctly handle LazyLLMDataset instances, ensuring proper statistical analysis when using multimodal lazy tokenization.
• Megatron Dataset Building Patch: A new patch has been introduced in swift/megatron/init.py to modify Megatron's build_train_valid_test_datasets function, ensuring correct handling of dataset providers for training, validation, and testing.
• Enhanced Logging for Evaluation Iterations: An informational log message has been added in swift/megatron/trainers/base.py to provide clearer visibility when args.eval_iters is set for streaming validation datasets.

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 introduces several bug fixes. In swift/llm/train/sft.py, _stat_dataset is updated to correctly handle LazyLLMDataset, which is crucial for multimodal models when lazy_tokenize is false. In swift/megatron/init.py, a patch is added for megatron.training.training.build_train_valid_test_datasets to ensure train_valid_test_num_samples is correctly passed, fixing a potential issue in the dataset building process for Megatron-based training. Finally, in swift/megatron/trainers/base.py, a log statement is added to improve visibility into the eval_iters setting. The changes are well-targeted and correct. The code quality is good, and I have no further suggestions.