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

Unfreezing weights during training #83

Open
amoskalev opened this issue May 5, 2020 · 4 comments
Open

Unfreezing weights during training #83

amoskalev opened this issue May 5, 2020 · 4 comments

Comments

@amoskalev
Copy link

Hi, thanks for your work,

Can you explain why do you have a training procedure with a gradual weights unfreezing? Why do you never train the first layer? What was the motivation?

Thanks,

@amoskalev
Copy link
Author

Hi, @JudasDie @Sy-Zhang @zhangliliang @HYPJUDY @ZhichengHuang can you please clarify the weight unfreezing procedure?

Thanks

@JudasDie
Copy link
Contributor

Hi, @JudasDie @Sy-Zhang @zhangliliang @HYPJUDY @ZhichengHuang can you please clarify the weight unfreezing procedure?

Thanks

Hi, sorry for my late replay. 1) Gradual weights unfreezing is faster than unfixing all layers at the beginning of training. 2) Of course you could train the first layer. However, in my experience, it doesn't bring more gains.

@ISosnovik
Copy link

@JudasDie Is it possible then to change the number of epochs or the number of steps or the learning rate, such that the same result is achieved faster?

@ISosnovik
Copy link

@JudasDie

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

3 participants