chore: upgrade torch 2.8.0

[zhyncs]

Motivation

Modifications

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://slack.sglang.ai to discuss your PR.

[gemini-code-assist]

Summary of Changes

Hello @zhyncs, 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 updated the project's dependencies to use a newer version of PyTorch. This involves bumping the minimum required Python version and adjusting the installation script for CI to ensure compatibility with the new PyTorch version.

Highlights

• Python Version Requirement: The minimum required Python version for the project has been updated from 3.8 to 3.9 in the pyproject.toml file.
• PyTorch Dependency Upgrade: The core 'torch' and 'torchaudio' dependencies have been upgraded from version 2.7.1 to 2.8.0 for both the 'srt' and 'blackwell' environments.
• torchvision Dependency Adjustment: The 'torchvision' dependency for the 'srt' environment no longer specifies a fixed version, providing more flexibility for its resolution.
• CI Installation Script Update: The CI installation script has been modified to include an extra index URL (https://download.pytorch.org/whl/test/cu126) when installing the 'python[dev]' package, which is likely necessary to correctly resolve and install the new PyTorch version in the CI environment.

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 upgrades PyTorch from version 2.7.1 to 2.8.0. The changes primarily involve updating the pyproject.toml file with the new dependency versions and modifying the CI script to use a test channel for PyTorch installation. My review identifies a couple of issues in pyproject.toml related to torchvision dependencies. Specifically, torchvision is unpinned in one dependency group and pinned to a likely incompatible version in another. These inconsistencies could lead to build failures or runtime errors and should be addressed. Additionally, I noticed that the documentation in docs/start/install.md still refers to the old torch version and should be updated to reflect this upgrade.

[gemini-code-assist]

torchvision is unpinned, which can lead to non-reproducible builds. It's recommended to pin it to a specific version compatible with torch==2.8.0. For example, torchvision==0.23.0 seems to be available on the PyTorch test channel. This would also make it consistent with the blackwell dependency group which has a pinned version (though that one also needs an update).

"torchvision==0.23.0",

[gemini-code-assist]

While torch and torchaudio are upgraded to 2.8.0, torchvision on the next line is kept at 0.22.1, which is likely incompatible. This can cause installation or runtime errors. Please update torchvision to a compatible version. This is also inconsistent with the srt dependency group where torchvision is unpinned.

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[wuz15]

@zhyncs Did you do the test before change it?

[EduardDurech]

@zhyncs @wuz15 doesn't work except with #8874 , could you please merge? It's been a week