Closed digantamisra98 closed 3 years ago
Reference: https://www.bmvc2020-conference.com/assets/papers/0928.pdf I didn't implement the backward pass using an autograd function like in the case of swish because of the internal stable differentiation of thresholded softplus.
Reference: https://www.bmvc2020-conference.com/assets/papers/0928.pdf I didn't implement the backward pass using an autograd function like in the case of swish because of the internal stable differentiation of thresholded softplus.