I would like to train two models using a single optimizer (and its state). The idea would be to transfer the model1.optimizer state to the second model.
I have been t