Weight sharing #62
Description
Hi,
Can you explain why multiply 0.5 from gradient ?
# Average the gradients
for p in D_shared.parameters():
p.grad.data = 0.5 * p.grad.data
{{ message }}
Hi,
Can you explain why multiply 0.5 from gradient ?
# Average the gradients
for p in D_shared.parameters():
p.grad.data = 0.5 * p.grad.data