CUDA: app option to compile without FlashAttention (#12025)
This commit is contained in:
parent
36c258ee92
commit
a28e0d5eb1
13 changed files with 46 additions and 31 deletions
|
@ -3203,7 +3203,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
|
|||
case GGML_OP_FLASH_ATTN_EXT: {
|
||||
#ifndef FLASH_ATTN_AVAILABLE
|
||||
return false;
|
||||
#endif
|
||||
#endif // FLASH_ATTN_AVAILABLE
|
||||
if (op->src[1]->type == GGML_TYPE_BF16 || op->src[2]->type == GGML_TYPE_BF16) {
|
||||
return false;
|
||||
}
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue