Fix resume offline train logic. Add loading optimizer state

[hanq-moreh]

Motivation

Add loading optimizer state when resume training.
Loading the optimizer state when resuming training is required to restore momentum, learning rate schedules, and other optimizer-specific statistics. Without it, the optimizer behaves as if starting from scratch.

Modifications

Related Issues

Accuracy Test

Benchmark & Profiling

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://sgl-fru7574.slack.com/archives/C09784E3EN6 to discuss your PR.

[gemini-code-assist]

Summary of Changes

Hello @hanq-moreh, 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 critical aspect of robust deep learning training by enhancing the resume logic for offline training. It ensures that when training is restarted from a checkpoint, not only are the model weights loaded, but also the optimizer's internal state and the correct training epoch and global step. This prevents the optimizer from behaving as if it's starting from scratch, thereby maintaining the integrity of the training process and improving convergence efficiency.

Highlights

• Optimizer State Loading: The pull request introduces functionality to load the optimizer's state when resuming training, which is crucial for restoring momentum, learning rate schedules, and other optimizer-specific statistics to ensure continuous and effective training.
• Dynamic Epoch Resumption: The training process now intelligently determines the correct starting epoch by parsing checkpoint directory names and the loaded training state, preventing redundant training and ensuring seamless continuation.
• Global Step Persistence: The global_step is now saved as part of the training state, allowing for more precise tracking and resumption of training progress beyond just the epoch number.

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 logic to resume offline training from a checkpoint, which includes loading the optimizer state and the global step. The implementation is solid and handles cases where older checkpoints might be missing the new information. I have one suggestion to improve the clarity of the fallback logic for calculating global_step.