Add out_sharding argument to Embed layer call

[samanklesaria]

This PR adds an out_sharding argument to Embed layers. Because jnp.take (which is what Embed wraps) does not provide an out_sharding argument itself, we handle this with auto_axes. The PR includes tests that this works as expected, and that without the out_sharding argument, mismatched sharding for arrays and indices is expected.

[gemini-code-assist]

Summary of Changes

Hello @samanklesaria, 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 significantly enhances the Embed layer within Flax's NNX module by introducing an out_sharding argument to its __call__ method. This addition provides users with fine-grained control over the sharding of the output array, which is critical for optimizing performance in distributed computing environments. The implementation cleverly leverages jax.sharding.auto_axes to overcome the inherent limitation of jnp.take not supporting output sharding directly, ensuring that the specified sharding is correctly applied. Comprehensive tests have been included to verify the functionality and ensure robust behavior under various sharding configurations.

Highlights

• Embed Layer Enhancement: The Embed layer's __call__ method now accepts an out_sharding argument, allowing explicit control over the output array's sharding.
• Sharding Implementation: Since jnp.take does not natively support out_sharding, jax.sharding.auto_axes is utilized to apply the specified output sharding to the result of the embedding operation.
• New Test Coverage: A new test case, test_out_sharding_embed, has been added to tests/nnx/spmd_test.py to validate the correct behavior of the out_sharding argument, including expected errors for sharding mismatches and successful sharding application.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

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 correctly adds an out_sharding argument to the Embed layer's __call__ method. The use of jax.sharding.auto_axes is appropriate since jnp.take does not natively support this argument. The accompanying test effectively validates the new functionality, including the expected failure case with mismatched sharding and the successful case with out_sharding. I've included a couple of suggestions to improve code readability and test robustness.