]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
ggml : drop support for QK_K=64 (#7473)
authorGeorgi Gerganov <redacted>
Thu, 23 May 2024 07:00:21 +0000 (10:00 +0300)
committerGitHub <redacted>
Thu, 23 May 2024 07:00:21 +0000 (10:00 +0300)
commite84b71c2c6da6e69c8f815168ea836f9716a325e
treeaa5c046cc5f0e9e953de5329412a753d82e60589
parent1b1e27cb49158123ef4902aa41eb368c9e76e6a1
ggml : drop support for QK_K=64 (#7473)

* ggml : drop support for QK_K=64

ggml-ci

* opencl : restore QK_K=256 define
16 files changed:
CMakeLists.txt
Makefile
ci/run.sh
ggml-common.h
ggml-cuda/convert.cu
ggml-cuda/dmmv.cu
ggml-cuda/mmq.cu
ggml-cuda/vecdotq.cuh
ggml-metal.m
ggml-metal.metal
ggml-opencl.cpp
ggml-quants.c
ggml-sycl.cpp
ggml.c
gguf-py/gguf/constants.py
llama.cpp