Closed ErenBalatkan closed 5 months ago
the network behaved like a Sparse net if (nempty=True). It should perform convolution on non-empty as well (nempty=False) for reconstruction.
Yes, your understanding and usage is correct. If (nempty=True), ocnn
behaves exactly the same with other Sparse nets which use Hash Tables.
If nempty=True, the tensor contains only features in non-empty octree nodes; otherwise, the tensor contains features in both non-empty and empty octree nodes. The size of the tensors is different. This is the reason for the error.
Thanks for the answer!
Greetings,
Thanks for the incredible work!
I'm experimenting with a setting where I would like to first downscale input voxel grid to /8 and then perform shape reconstruction from that point. For the encoding layer, I would prefer if network behaved like a Sparse net (nempty=True) but during upscaling, since I want it to perform reconstruction it should perform convolution on non-empty as well (nempty=False).
I have tried to test this approach with a simple script, but it is giving me shape error.