OrgLion-ML
/
darknet
mirror of https://github.com/AlexeyAB/darknet.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Added __float2half_rn() and __half2float()

This commit is contained in:
AlexeyAB 2018-02-22 23:52:43 +03:00
parent f558d5c39c
commit b2b5756d86
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/convolutional_kernels.cu
View File

@ -77,7 +77,7 @@ void binarize_weights_gpu(float *weights, int n, int size, float *binary)
__global__ void cuda_f32_to_f16(float* input_f32, size_t size, half *output_f16)
{
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx < size) output_f16[idx] = input_f32[idx];
if (idx < size) output_f16[idx] = __float2half_rn(input_f32[idx]);
}
void cuda_convert_f32_to_f16(float* input_f32, size_t size, half *output_f16) {
@ -87,7 +87,7 @@ void cuda_convert_f32_to_f16(float* input_f32, size_t size, half *output_f16) {
__global__ void cuda_f16_to_f32(half* input_f16, size_t size, float *output_f32)
{
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx < size) output_f32[idx] = input_f16[idx];
if (idx < size) output_f32[idx] = __half2float(input_f16[idx]);
}
void cuda_convert_f16_to_f32(half* input_f16, size_t size, float *output_f32) {