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

restore some modifications

Browse source
This commit is contained in:
shanshan shen 2024-11-26 08:49:26 +00:00
parent e05a398fb3
commit 33fd470550
1 changed files with 0 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

1
ggml/src/ggml-cann/aclnn_ops.cpp
View file

@ -2427,7 +2427,6 @@ static void aclnn_mat_mul(ggml_backend_cann_context& ctx, aclTensor* acl_input,
aclTensor* acl_weight, aclTensor* acl_dst) { aclTensor* acl_weight, aclTensor* acl_dst) {
int8_t cube_math_type = 1; // ALLOW_FP32_DOWN_PRECISION, when input is int8_t cube_math_type = 1; // ALLOW_FP32_DOWN_PRECISION, when input is
// fp32, atlas a2 will transpose it to HFLOAT32. // fp32, atlas a2 will transpose it to HFLOAT32.
uint64_t workspaceSize = 0; uint64_t workspaceSize = 0;
aclOpExecutor* executor; aclOpExecutor* executor;
void* workspaceAddr = nullptr; void* workspaceAddr = nullptr;