Skip to content

Conversation

http403
Copy link

@http403 http403 commented May 12, 2025

This PR add new documentation for setting up tabbyAPI in WSL2 with an AMD GPU.

Working environment:

  • Host OS: Windows 11 Build 26100
  • WSL version: 2.5.7.0
  • Ubuntu version: 24.04
  • Kernel version: 5.15.167.4-microsoft-standard-WSL2+ (self-compiled)
  • tabbyAPI commit: b555eeb
  • GPU: AMD 7900XTX
  • AMD driver version: 25.3.1

Know limitations:

  • Torch HSA runtime must be replaced

I can provide the exported WSL2 machine if anyone need (~18GB).

@kingbri1
Copy link
Member

This looks good overall, but should we be telling users how to install WSL? I'd rather link to the official docs in case something changes in the future and that not all Windows installations are equal.

@http403
Copy link
Author

http403 commented May 17, 2025

Thanks for the suggestion. I have update the respective install instructions to their respective official websites.

Migrate links to Markdown links for conciseness.
@http403
Copy link
Author

http403 commented May 19, 2025

Thanks for the edits. Sorry for the bad writing skill.

Comment on lines +63 to +66
3. Update your WSL system
```bash
sudo apt update && sudo apt upgrade -y
```
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If the docs already refers to official websites for instruction, should the run and configure (update) also delegates to Ubuntu's docs instead?

@http403 http403 marked this pull request as draft May 21, 2025 21:14
@http403
Copy link
Author

http403 commented May 21, 2025

I just found AMD's doc that they only support PyTorch 2.4.0 on WSL. More tests has to be done. I have converted this PR to a draft for now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants