llama-cpp-bindings-sys 0.7.0

Low level bindings to llama.cpp
Documentation 
1
2
3
4
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);