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

CUDA: remove incorrect precision check

Browse source
This commit is contained in:
Johannes Gäßler 2024-05-22 09:28:32 +02:00
parent 201cc11afa
commit 5cf3c42412
1 changed files with 0 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ggml-cuda/fattn-tile-f32.cu
View file

@ -286,9 +286,6 @@ void ggml_cuda_flash_attn_ext_tile_f32(ggml_backend_cuda_context & ctx, ggml_ten
const ggml_tensor * KQV = dst;
const ggml_tensor * Q = dst->src[0];
const int32_t precision = KQV->op_params[2];
GGML_ASSERT(precision == GGML_PREC_DEFAULT);
if (Q->ne[1] <= 16) {
constexpr int cols_per_block = 16;
constexpr int parallel_blocks = 4;