PyTorch Forums
Zero grad on single parameter
sparseinference
(Peter Cavén)
March 17, 2019, 9:08pm
2
The gradient will be computed when you call:
a.backward()
1 Like
show post in topic