Key error in transposed convolution
akshay-antony opened this issue · comments
Is there an existing issue for this?
- I have searched the existing issues
Current Behavior
I am making a unet kind of model, with skip connections with 3 encoders (stride2), and 3 decoder layers. But it seems the transposed convolution of the last decoder layer is giving the following error:
main_path = self.relu(self.bn1(self.transposed_conv1(x)))
File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "torchsparse/nn/modules/conv.pyx", line 99, in torchsparse.nn.modules.conv.Conv3d.forward
File "torchsparse/nn/functional/conv/conv.pyx", line 148, in torchsparse.nn.functional.conv.conv.conv3d
KeyError: (1, 1, 1)
All the other transposed convolution works fine
Expected Behavior
No response
Environment
- GCC:
- NVCC:
- PyTorch:
- PyTorch CUDA:
- TorchSparse:Anything else?
No response
Could you please provide a short code snippet to reproduce this error? Thank you!