Skip to content

sync : ggml + llama.cpp #2455

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
Merged
merged 15 commits into from
Oct 5, 2024
Merged

sync : ggml + llama.cpp #2455

merged 15 commits into from
Oct 5, 2024

Conversation

ggerganov
Copy link
Member

@ggerganov ggerganov commented Oct 5, 2024 •
edited
Loading

Also:

  • Revert the recent Mel related changes in order to reduce complexity
  • Zero-init Metal contexts, sometimes causing seg-faults on program exit when having empty compute buffers (e.g. CoreML encoder + Metal decoder)

ggerganov and others added 15 commits October 5, 2024 13:09
@ggerganov
scripts : sync ggml-backend.cpp
f97be1b
@JohannesGaessler @ggerganov
ggml: refactor cross entropy loss CPU impl. (ggml/976)
1a5d63d
@JohannesGaessler @ggerganov
ggml/ex: calculate accuracy in graph, adapt MNIST (ggml/980)
8fdc942
@rgerganov @0cc4m @ggerganov
vulkan : do not use tensor->extra (llama/9407)
d6f1bd9 
* vulkan : do not use tensor->extra

This patch allows using the Vulkan backend with the RPC backend as
tensor->extra is no longer used.

Ref: #8536

* Adapt GGML_VULKAN_CHECK_RESULTS to extra removal (llama/2)

---------

Co-authored-by: 0cc4m <picard12@live.de>
@Alcpz @ggerganov
Initial cmake support of SYCL for AMD GPUs (llama/9658)
1163865 
sycl: initial cmake support of SYCL for AMD GPUs
@slaren @JohannesGaessler @ggerganov
ggml-backend : add device and backend reg interfaces (llama/9707)
3d86dc4 
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
@OuadiElfarouki @ggerganov
Fixed dequant precision issues in Q4_1 and Q5_1 (llama/9711)
84cc6c8
@slaren @ggerganov
ggml-backend : add device and backend reg interfaces (llama/9707)
cd78b88 
Also:

- metal : fix compute pass descriptor autorelease crash
- ggml-backend : add device description to CPU backend
- ggml: unify backend logging mechanism
@slaren @ggerganov
ggml : fixes after sync (ggml/983)
66225ab 
ggml : remove test-backend-buffer

ggml : fix CUDA build warnings
@danbev @ggerganov
ggml : fix typo in example usage ggml_gallocr_new (ggml/984)
98a5408
@ggerganov
whisper : adapt to latest ggml (skip) (#0)
f287437
@ggerganov
whisper : revert mel-related changes (#0)
d9a6ba3 
too much extra logic and complexity for small benefit
@ggerganov
metal : zero-init buffer contexts (#0)
7a2784e
@ggerganov
objc : fix build
0c0247e
@ggerganov
whisper : zero-out the KV cache upon clear (#2445)
47de411
@ggerganov ggerganov merged commit 847f94f into master Oct 5, 2024
37 checks passed
@ggerganov ggerganov deleted the sync branch October 5, 2024 12:23
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.
7 participants
@ggerganov @JohannesGaessler @rgerganov @Alcpz @slaren @OuadiElfarouki @danbev