Key error when using deconv

[akshay-antony]

/lib/python3.7/site-packages/torchsparse/nn/functional/conv.py", line 134, in conv3d kmap = input.kmaps[(tensor_stride, kernel_size, stride, dilation)] KeyError: ((8, 8, 2), (3, 3, 3), (2, 2, 2), (1, 1, 1))

I am writing Cylinder3d using torchsparse, but a stride 2 deconv is causing key error. And is there a way to share the coordinate mapping Minkowski or spconv?

[zhijian-liu]

@akshay-antony, thank you for reporting this issue! Could you please provide a short code snippet to reproduce this bug?

[zhijian-liu]

Since TorchSparse has been upgraded to v2.1.0, could you please attempt to install the latest version? I will now close this issue, but please don't hesitate to reopen it if the problem persists.