Skip to content

fix initial_lr is not specified in param_groups[0] when reuming#14

Open
LiUzHiAn wants to merge 1 commit into
akamaster:masterfrom
LiUzHiAn:resume_fix
Open

fix initial_lr is not specified in param_groups[0] when reuming#14
LiUzHiAn wants to merge 1 commit into
akamaster:masterfrom
LiUzHiAn:resume_fix

Conversation

@LiUzHiAn

Copy link
Copy Markdown

Hi,

When I tried to resume the training process, I was told that 'initial_lr' is not specified in param_groups[0] when resuming an optimizer. So, I guess we have to feed this key to the optimizer. That is:

optimizer = torch.optim.SGD([{"params":model.parameters(),"initial_lr":args.lr}], lr=args.lr,
                                momentum=args.momentum,
                                weight_decay=args.weight_decay)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant