Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

fix YaRN for NeoX type #839

Closed
wants to merge 2 commits into from
Closed

fix YaRN for NeoX type #839

wants to merge 2 commits into from

Conversation

foldl
Copy link
Contributor

@foldl foldl commented May 27, 2024

This PR will fix #835.

@ggerganov
Copy link
Owner

Thanks! Do the GPU kernels need to be updated?

@foldl
Copy link
Contributor Author

foldl commented May 28, 2024

GPU backends:

  1. CUDA: looks Good.
  2. Kompute: not sure. Where is shaderop_rope_xx.h?
  3. Metal: Fixed.
  4. SYCL: looks Good.
  5. Vulkan: Fixed.

But, unfortunately, I don't have environments to test these GPU fixes.

@ggerganov
Copy link
Owner

Had top merge a few changes from llama.cpp and we have some conflicts - sorry about that.
After resolving these, could you try to extend test-backend-ops to exercise this code via ggml_rope_ext()?

@foldl
Copy link
Contributor Author

foldl commented May 29, 2024

Sorry, but I don't think I have time to update test-backend-ops. Maybe later, or someone else could help on this.

@ggerganov ggerganov mentioned this pull request May 29, 2024
Merged
@ggerganov
Copy link
Owner

superseded by ggerganov/llama.cpp#7617

@ggerganov ggerganov closed this May 29, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

issues about YaRN
2 participants
@foldl @ggerganov