llama.cpp/ggml-cuda/im2col.cuh

6 lines
137 B
Plaintext

#include "common.cuh"
#define CUDA_IM2COL_BLOCK_SIZE 256
void ggml_cuda_op_im2col(ggml_backend_cuda_context & ctx, ggml_tensor * dst);