diff options
Diffstat (limited to 'llama.cpp/ggml/src/ggml-cuda/fattn.cuh')
| -rw-r--r-- | llama.cpp/ggml/src/ggml-cuda/fattn.cuh | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/llama.cpp/ggml/src/ggml-cuda/fattn.cuh b/llama.cpp/ggml/src/ggml-cuda/fattn.cuh new file mode 100644 index 0000000..78705d5 --- /dev/null +++ b/llama.cpp/ggml/src/ggml-cuda/fattn.cuh @@ -0,0 +1,5 @@ +#include "common.cuh" + +void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst); + +bool ggml_cuda_flash_attn_ext_supported(int device, const ggml_tensor * dst); |