summaryrefslogtreecommitdiff
path: root/ggml-cuda/convert.cuh
blob: db34c0be96449706c2b3e6d441ffbd9a54b312da (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
#include "common.cuh"

#define CUDA_DEQUANTIZE_BLOCK_SIZE 256

template<typename T>
using to_t_cuda_t = void (*)(const void * __restrict__ x, T * __restrict__ y, int k, cudaStream_t stream);

typedef to_t_cuda_t<float> to_fp32_cuda_t;
typedef to_t_cuda_t<half> to_fp16_cuda_t;

to_fp16_cuda_t ggml_get_to_fp16_cuda(ggml_type type);

to_fp32_cuda_t ggml_get_to_fp32_cuda(ggml_type type);