out_channels for 3D binary classification with SEnets

[tommydino93]

Hi All :)

I am performing 3D classification using densenet. However, I would be curious to try also the SENets. If I'm not mistaken, currently all the SE nets inherit from SENet which is 2D. Are you planning to extend these networks also for 3D?

Thanks a lot in advance!
Tommaso

[Nic-Ma]

Hi @tommydino93 ,

I think all the SENet in MONAI can support 3D data, @yiheng-wang-nv could you please help confirm it?

Thanks in advance.

[tommydino93]

Hi @Nic-Ma ,

Actually, I expressed myself wrong. You are right. I can put spatial_dims=3.

The issue is more on how to set the out_channels argument. I am performing binary classification, so I would like to set out_channels=2 (as I do for DenseNet), but with SENet I get Unexpected argument for out_channels . What would be the equivalent argument?

Thanks again and sorry for the inaccuracy

[Nic-Ma]

Maybe you can try the num_classes arg?

Thanks.

[tommydino93]

Yes, it works! Thank you very much for your time :)