[Relax][PyTorch] Fix _slice and _expand for dynamic shapes in PyTorch ExportedProgram frontend

[mshr-h]

Fixes two issues when translating PyTorch models with dynamic shapes:

1. _slice: Resolve fx.Node references in start/end/step arguments and detect identity slices where the symbolic end equals the tensor dimension (avoids redundant strided_slice ops).

2. _expand: Fall back to FX node metadata when shape_of() returns None for tensors with unknown shapes.

[gemini-code-assist]

Summary of Changes

Hello, 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 enhances the PyTorch ExportedProgram frontend's ability to correctly handle dynamic shapes by improving the translation of _slice and _expand operations. These changes ensure that symbolic dimensions are properly resolved and that identity operations are optimized away, leading to more accurate and efficient Relax IR generation for dynamic PyTorch models.

Highlights

• Dynamic Shape Handling in _slice: Resolved fx.Node references in start, end, and step arguments for _slice operations, enabling correct interpretation of symbolic dimensions. Additionally, implemented logic to detect and elide identity slices where the symbolic end value matches the tensor's dimension, preventing redundant strided_slice operations.
• Dynamic Shape Handling in _expand: Introduced a fallback mechanism in the _expand function to retrieve shape information from FX node metadata (FakeTensor) when shape_of() returns None for tensors with unknown shapes, improving robustness for dynamic shape scenarios.

🧠 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.

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 addresses issues with dynamic shapes in the PyTorch frontend for _slice and _expand operations. The _slice operator is updated to resolve symbolic arguments and to identify and elide identity slices with dynamic shapes, which avoids creating redundant strided_slice operations. The _expand operator is enhanced with a fallback mechanism to use shape information from FX node metadata when it's not available at the Relax level. The changes are accompanied by new tests that verify the correct handling of symbolic shapes for slicing.

[mshr-h]

cc @tlopex @guan404ming