Skip to content
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.

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

Training settings #18

Open
dwangxxx opened this issue Oct 26, 2020 · 1 comment
Open

Training settings #18

dwangxxx opened this issue Oct 26, 2020 · 1 comment

Comments

@dwangxxx
Copy link

Hello, I've read your paper recently. In this paper, you mentioned that the training epoch is set as 2, 5 and 2, but the source code is set 150. I want to know which is correct? Thank you!

@dwangxxx
Copy link
Author

Hello, sorry to bother again. In this paper, you show the results finetuned on the COCOstuff dataset. I want to know how to finetune the model on the COCOstuff dataset for ImageNet dataset. I don't know if I'm right, but I thought of the following steps:
(1) Train the model on the ImageNet dataset.
(2) Finetune the model from (1) on the COCOStuff dataset with a smaller learning rate.
Maybe I was wrong, but can you provide me some detailed information, such as finetune learning rate and some other parameters?

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

No branches or pull requests

1 participant