Files
llama.cpp/ggml/src/ggml-cuda/fattn.cuh
2025-08-20 23:14:14 +02:00

6 lines
185 B
Plaintext

#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);