Repository navigation
padding="same" and groups=in_channels in 3D convolution not supported #36
Description
Activity
- changed the title
[-]AssertionError: padding="same" and groups=[/-][+]AssertionError: padding="same" and groups=in_channels in 3D convolution not supported[/+]on Apr 25, 2024 - changed the title
[-]AssertionError: padding="same" and groups=in_channels in 3D convolution not supported[/-][+]padding="same" and groups=in_channels in 3D convolution not supported[/+]on Apr 25, 2024 Thanks for opening the issue! Could you share a minimal snippet of code to reproduce the issue?
I have tried to make a 3D convolutional layer with the following parameters:
padding='same'andgroups='in_channels'and l get anAssertionError.Either withtltorch.FactorizedConv()ortltorch.FactorizedConv.from_conv()It seems that
padding='same'withgroupsis not supported in tensorly.padding='same'in pytorch pytorch doc on paddinggroups= in_channelsin pytorch groups in convThank you for your help.