PyTorch Forums
Gradient calculation when module used multiple times (e.g. siamese network)
smth
June 6, 2017, 1:41pm
2
the gradients are summed (i.e. accumulated).
3 Likes
show post in topic