gungui98 / Pytorch-Depthwise-Conv3d

cuda implementation of depthwise conv3d

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Accelerated 3D Depthwise Convolution

This is seperate repo of my pull request (Accelerated 3D Depthwise Convolution), which is part of Pytorch 1.9. This repo aim to support other people want to use the module without upgrade to latest cudnn or pytorch.



  • Pytorch >= 1.6
  • Python3
python install


import torch
from depthwise_conv3d import DepthwiseConv3d

dtype = torch.float
conv = DepthwiseConv3d(2, 2, kernel_size=3, groups=2).to("cuda", dtype)
input = torch.randn(2, 2, 6, 6, 6, device="cuda", dtype=dtype).div_(2).requires_grad_()
output = conv(input)


cuda implementation of depthwise conv3d

License:MIT License


Language:Cuda 76.5%Language:Python 20.6%Language:C++ 2.8%