Update callbacks.py

[franklin5]

AttributeError: 'Functional' object has no attribute '_distribution_strategy'.

distribute_strategy is a property of the model

[gemini-code-assist]

Summary of Changes

Hello @franklin5, 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!

I've addressed an AttributeError that occurred when accessing the _distribution_strategy attribute on a model within tf_keras/callbacks.py. The issue stemmed from attempting to access a private attribute (_distribution_strategy) that does not exist, whereas the correct public property distribute_strategy should be used. This change ensures that the distribution strategy is correctly retrieved, resolving the runtime error.

Highlights

• Bug Fix: Corrected an AttributeError in tf_keras/callbacks.py by updating the access of the model's distribution strategy from self.model._distribution_strategy to self.model.distribute_strategy. This aligns with the public API for accessing the distribution strategy.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 or fill out our survey to provide feedback.

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 an AttributeError in the BackupAndRestore callback by using the public distribute_strategy property instead of the private _distribution_strategy attribute. While the change is in the right direction, it introduces a side effect where a NotImplementedError could be raised for users not using any distribution strategy. I've provided a suggestion to refine the logic to correctly handle the default strategy case, ensuring the callback behaves as expected for all users.

[gemini-code-assist]

While this change correctly uses the public distribute_strategy property and fixes the AttributeError, it introduces a potential issue. The property self.model.distribute_strategy falls back to tf.distribute.get_strategy(), which returns a DefaultDistributionStrategy instance if no other strategy is active. This means the property is never None, and the if condition will always be evaluated.

Since DefaultDistributionStrategy is likely not in self._supported_strategies, this will cause a NotImplementedError to be raised for users who are not using any distribution strategy, which is not the intended behavior.

The logic should be adjusted to explicitly allow the DefaultDistributionStrategy to pass through without raising an error. This ensures the callback only validates strategies when a non-default one is actually in use.

        if not isinstance(
            self.model.distribute_strategy,
            (tf.distribute.DefaultDistributionStrategy, *self._supported_strategies),
        ):