Adding Tensor_layout for Tensor parallelism for Autosharding

[buildwithsuhana]

This change introduces core building blocks for tensor parallelism by adding two key components.

First, it adds crucial collective operations, all_reduce and all_gather, to the JAX backend. These allow multiple devices to synchronize data by summing tensors (like gradients) or gathering individual slices back into a full tensor. Second, it adds the high-level tensor sharding logic (split_tensor_for_parallelism), which uses ops.array_split to intelligently slice large tensors, even unevenly, for distribution across devices. New tests confirm this new parallel logic, including the uneven splitting, works as expected.

The tests on this PR will pass after the PR #21697 gets merged

[gemini-code-assist]

Summary of Changes

Hello @buildwithsuhana, 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 establishes foundational components for tensor parallelism within the JAX backend, crucial for Autosharding. It provides core collective communication primitives like all_reduce and all_gather and introduces a flexible tensor splitting utility, split_tensor_for_parallelism, designed to efficiently distribute tensors across multiple devices, even when uneven splitting is required.

Highlights

• JAX Collective Operations: Introduced all_reduce (sum, mean) and all_gather functions to the JAX backend for inter-device communication, essential for distributed computing.
• Tensor Sharding Logic: Added split_tensor_for_parallelism to intelligently slice tensors, including support for uneven distributions, for efficient device parallelism.
• Comprehensive Testing: Included new test cases to validate the correct functionality of both the collective operations and the tensor splitting logic, ensuring robustness.

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 foundational components for tensor parallelism in Keras, specifically for the JAX backend. It adds all_reduce and all_gather collective operations, which are essential for distributed computations. Additionally, it provides a split_tensor_for_parallelism utility for sharding tensors across devices. The changes are well-tested, covering both even and uneven tensor splitting. My review includes a few suggestions to improve documentation accuracy and code simplicity, and to align with the repository's style guide regarding docstring examples.