Files
llama.cpp-mtp-turboquant/ggml/src/ggml-cuda/fattn-tile.cuh
T
2025-09-07 00:26:28 +02:00

4 lines
111 B
Plaintext

#include "common.cuh"
void ggml_cuda_flash_attn_ext_tile(ggml_backend_cuda_context & ctx, ggml_tensor * dst);