Open ahkarami opened 6 years ago
@ahkarami Thank you.
The models that we are using don't use either dropout or 3d convolution unfortunately so it's not a direct priority.
Although it's quite easy to do for 3d convolution for example: https://github.com/warmspringwinds/pytorch-cpp/blob/master/src/pytorch.cpp#L375
You should just use different function instead of SpatialDilatedConvolution_updateOutput
.
If you have interest in these I can guide you on how to do it.
Thank you.
Dear @warmspringwinds, Thank you for your time and response. About implementing 3d convolution and other layers, maybe in the near future I will do that. By the way, I think your project was really fantastic and useful. I hope that you continue to progress it. If you want to know about my reasons, you can see the below link: [Feature Request] Deployment in c++ project Best
@ahkarami I am glad you liked it :)
Yes, the project is alive an developing -- our focus is just directed towards 2d data so far. We will add support for 3d soonish.
Dear @warmspringwinds, Your project is fantastic. Do you have nay plan for continuing it? For example by implementing other PyTorch's layers (e.g., nn.Dropout2d, nn.Conv3d)?