IBM/simulai

Is the SciPyInterface really communicating gradients to the torch.nn.Module objects ?

Opened this issue · 0 comments

It seems that optimizers like L-BFGS-B are not being really effective.