whisper.cpp/ggml-cuda/pad.cuh

6 lines
131 B
Plaintext
Raw Normal View History

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