search

wavefrontshaping / complexPyTorch

A high-level toolbox for using complex valued neural networks in PyTorch
MIT License
604 stars 146 forks source link

torchinfo.summary() is not able to executable #31

Open Weijie-ZHAO opened 1 year ago

Weijie-ZHAO commented 1 year ago

image

Weijie-ZHAO commented 1 year ago

Hi, thanks for your liabrary, but could you pls solve this problem when we run torchinfo.summary() to check each layer

Best!

wavefrontshaping commented 10 months ago

Hi,

Are you sure you specified the correct dtype? You should use complex64. It works for me with something like

torchinfo.summary(model, input_size=(1, 1, 28, 28), dtypes = [torch.complex64])