Error message: p.grad.data.mul_(1.0 / meta_batch_size)
AttributeError: 'NoneType' object has no attribute 'data'
For lines;
Average the accumulated gradients and optimize
for p in maml.parameters():
print(p)
p.grad.data.mul_(1.0 / meta_batch_size)
opt.step()
I encountered the error message in #387 and did as recommended, added the parameter allow_unused=True.
Would it be wrong if I changed the aforementioned lines to:
with torch.nograd():
for p in maml.parameters():
if p.grad is not None:
p.grad.data.mul(1.0 / meta_batch_size)
opt.step()
Error message: p.grad.data.mul_(1.0 / meta_batch_size) AttributeError: 'NoneType' object has no attribute 'data'
For lines;
Average the accumulated gradients and optimize
I encountered the error message in #387 and did as recommended, added the parameter allow_unused=True.
Would it be wrong if I changed the aforementioned lines to: with torch.nograd(): for p in maml.parameters(): if p.grad is not None: p.grad.data.mul(1.0 / meta_batch_size) opt.step()