-
-
Notifications
You must be signed in to change notification settings - Fork 2k
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.
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
feat: update whisper_cpp with CUBLAS, HIPBLAS, METAL, OPENBLAS, CLBLAST support #1302
Conversation
disabled WHISPER_METAL gpu acceleration, to make it work as before on metal, not causing regression. if build-apple works i would open a feature parity issue for whisper.cpp GPU on METAL support as reminder, but whisper.cpp would be up2date and gained GPU accel on CUDA, HIP |
tests-apple fail (no matter if WHISPER_METAL=1 or not). upstream ticket: ggerganov/whisper.cpp#1344
|
thanks @wuxxin - this is looking good from my side - let's wait for upstream and merge it once the CI is green |
✅ Deploy Preview for localai canceled.
|
pr ggerganov/whisper.cpp#1530 has been merged, commit hash updated to 1.15.1 |
@mudler pr ggerganov/whisper.cpp#1530 has been merged, commit hash updated to 1.15.1 |
mm seems it still fails @wuxxin |
Head branch was pushed to by a user without write access
👍 |
Description
This PR updates whisper_cpp to release 1.5 and adds GPU Support
i build and tested ROCM/HIPBLAS GPU support for whisper.
i did not test any other combination.
This is a draft. Please ignore if not fitting.
Notes for Reviewers
Signed commits