RuntimeError: The expanded size of the tensor (1) must match the existing size (64) at non-singleton dimension 3. Target sizes: [2, 1, 1, 1]. Tensor sizes: [2, 128, 64, 64] (inferExpandGeometry at /pytorch/aten/src/ATen/ExpandUtils.cpp:75)

[wangy9711]

Hey, thanks for your work.
I want to try your net, I did create the conda environment. But when I execute python demo.py:

I only have one GPU, RTX2080.
Waiting for your suggestions.

[heilaw]

It may have something to do with the compilation of the corner pooling layers. Were you using GCC 4.9.2 or above to compile the layers?

[wangy9711]

It may have something to do with the compilation of the corner pooling layers. Were you using GCC 4.9.2 or above to compile the layers?

My gcc version is 5.4.0, Ubuntu version is 16.04.

[owen800q]

It may have something to do with the compilation of the corner pooling layers. Were you using GCC 4.9.2 or above to compile the layers?

My gcc version is 5.4.0, Ubuntu version is 16.04.

It may have something to do with the compilation of the corner pooling layers. Were you using GCC 4.9.2 or above to compile the layers?

My gcc version is 5.4.0, Ubuntu version is 16.04.

I encountered this problem too before, it works perfectly when I upgraded it to ubuntu 18.04

[wangy9711]

I find the solution by following change <torch/torch.h> to <torch/extension.h>.
After did that, I can get the result.
Thanks!