whisper.cpp / ggml-cuda /template-instances /fattn-vec-f16-instance-hs128-q8_0-q5_0.cu
JohannesGaessler's picture
CUDA: refactor mmq, dmmv, mmvq (llama/7716)
849ff52
raw
history blame
179 Bytes
// This file has been autogenerated by generate_cu_files.py, do not edit manually.
#include "../fattn-vec-f16.cuh"
DECL_FATTN_VEC_F16_CASE(128, GGML_TYPE_Q8_0, GGML_TYPE_Q5_0);