Results
1
comments of
Q
> In the MLP model, I think in the last layer it should be F.log_softmax instead of softmax. > Otherwise, the NLL loss would return negative values. I agree with...