AvgPool2d in Downsample

Pritam-N / ParNet

Custom Implementation of Non-Deep Networks

MIT License

26 stars 9 forks source link

AvgPool2d in Downsample #2

Open JJZHK opened 2 years ago

JJZHK commented 2 years ago

In Downsample, I think it should be torch.nn.AvgPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1)), not torch.nn.AvgPool2d(kernel_size=(2,2)). otherwise x & y have wrong size, and x + y will throw exception.

Pritam-N commented 2 years ago

@JJZHK Based on the architecture in the paper, it has avgpool (2,2) and the code is working.