p20518467
/
InfiniTensor
forked from jiuyuan/InfiniTensor
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Conv Kernel FP16

This commit is contained in:
zhangyue207 2023-07-19 10:06:47 +08:00
parent 523946cb8b
commit 48322dbf27
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/kernels/cuda/conv_half.cc
View File

@ -33,7 +33,7 @@ struct ConvCuDnnPerfRecordObj : public PerfRecordObj {
using ConvCuDnnPerfRecord = Ref<ConvCuDnnPerfRecordObj>;
class convCudnnHalf : public Kernel {
class convCudnnFP16 : public Kernel {
static constexpr int N_ALGO = 8;
static constexpr int N_MODE = 2;
@ -294,7 +294,7 @@ class convCudnnHalf : public Kernel {
}
};
REGISTER_KERNEL(Device::CUDA, OpType::Conv, DataType::Float16, convCudnnHalf,
REGISTER_KERNEL(Device::CUDA, OpType::Conv, DataType::Float16, convCudnnFP16,
"Conv_cuDNN_CUDA_Float16");
REGISTER_CONSTRUCTOR(1, ConvCuDnnPerfRecordObj::from_json);