New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

losses appear to be 0 #194

Open

1YuChenyao opened this issue Aug 15, 2024 · 2 comments

1YuChenyao commented Aug 15, 2024

The learning rate has been reduced to 1e-4, and 2 pictures were run at 4090. Why do some losses appear to be 0 during the second phase of training?

OrangeSodahub commented Aug 22, 2024 •

edited

Loading

@1YuChenyao Yes, I have the same question as you. Did you figure it out? In my case, the planning.loss_collision is zero

OrangeSodahub mentioned this issue

Training loss keeps 0 #196

Open

Author

1YuChenyao commented Aug 22, 2024 via email

Sorry，I am not solve this question

---- Replied Message ---- | From | ***@***.***> | | Date | 08/22/2024 11:54 | | To | OpenDriveLab/UniAD ***@***.***> | | Cc | ycy ***@***.***>, Mention ***@***.***> | | Subject | Re: [OpenDriveLab/UniAD] losses appear to be 0 (Issue #194) | @1YuChenyao Yes, I have the same question as you. Did you figure it out? In my case, the planning.loss_collision is zero! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

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