Should backward function be written every time I add a model

[thtb]

Then it's very time consuming

[hwang595]

Hi @thtb, in PyTorch, you won't need to write a backward function for your model, the autograd library will handle it for you.

But if you're quoting the backward function I kept in LeNetSplit model and ResNetSplit model, please ignore them. Those are for trials of my another project. I will delete those blocks of code soon to avoid making people confused.

Hope this helps.