Open etienne87 opened 2 years ago
Hi, that means the batch sizes don't match, and the library doesn't know how to deal with that situation.
Since PyTorch's broadcasting rules are extensive, not all rules are supported yet.
I'll see what I can do about it in the upcoming changes in #18 with a much more modular implementation.
I am trying to apply koila lazy eval on a Unet3D.
This fails and outputs: