llama.cpp/ggml-cuda/upscale.cuh

6 lines
139 B
Plaintext

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