vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

allow iq quant

Browse source
This commit is contained in:
abhilash1910 2024-03-15 00:33:35 -07:00
parent 0af3ed733f
commit 87e5c86686
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml-sycl.cpp
View file

@ -15498,7 +15498,7 @@ static void ggml_sycl_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
#ifdef GGML_SYCL_FORCE_DMMV #ifdef GGML_SYCL_FORCE_DMMV
const bool use_mul_mat_vec_q = false; const bool use_mul_mat_vec_q = false;
#else #else
const bool use_mul_mat_vec_q = min_compute_capability >= VER_4VEC && ggml_is_quantized(src0->type) && ggml_nrows(src1) == 1; const bool use_mul_mat_vec_q = min_compute_capability >= VER_4VEC && ggml_is_quantized(src0->type);
#endif // GGML_SYCL_FORCE_DMMV #endif // GGML_SYCL_FORCE_DMMV
if (use_mul_mat_vec_q) { if (use_mul_mat_vec_q) {