Some confusions about torch.optim.Adam().step()'s principle

sry i forgot the optimizer i choose is Adam…not SGD