diff options
author | Johannes Gäßler <johannesg@5d6.de> | 2024-06-05 16:53:00 +0200 |
---|---|---|
committer | GitHub <noreply@github.com> | 2024-06-05 16:53:00 +0200 |
commit | 7d1a378b8fb266782d9248538a661405aad80768 (patch) | |
tree | 7ce459a4c5a85e75f75825772124aedc3bb54b7f /ggml-cuda/template-instances/mmq-instance-q2_k.cu | |
parent | 2b3389677a833cee0880226533a1768b1a9508d2 (diff) |
CUDA: refactor mmq, dmmv, mmvq (#7716)
* CUDA: refactor mmq, dmmv, mmvq
* fix out-of-bounds write
* struct for qk, qr, qi
* fix cmake build
* mmq_type_traits
Diffstat (limited to 'ggml-cuda/template-instances/mmq-instance-q2_k.cu')
-rw-r--r-- | ggml-cuda/template-instances/mmq-instance-q2_k.cu | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/ggml-cuda/template-instances/mmq-instance-q2_k.cu b/ggml-cuda/template-instances/mmq-instance-q2_k.cu new file mode 100644 index 00000000..6415369d --- /dev/null +++ b/ggml-cuda/template-instances/mmq-instance-q2_k.cu @@ -0,0 +1,5 @@ +// This file has been autogenerated by generate_cu_files.py, do not edit manually. + +#include "../mmq.cuh" + +DECL_MMQ_CASE(GGML_TYPE_Q2_K); |