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

Implementing the Decayed Adagrad optimizer operator #4645

Merged
merged 4 commits into from
Oct 12, 2017

Conversation

kexinzhao
Copy link
Contributor

Fixes #4643

Formula for decayed Adagrad operator:

moment_out = decay * moment + (1 - decay) * grad * grad
param_out = param - learning_rate * grad / (sqrt(moment_out) + epsilon)


Decayed Adagrad

moment_out = decay * moment + (1 - decay) * grad * grad
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

does there have an article about this optimize algorithm?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I cannot find published article about it.
Just following the formula in http://doc.paddlepaddle.org/develop/doc/api/v2/config/optimizer.html#decayedadagrad

Copy link
Member

@jacquesqiao jacquesqiao left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@kexinzhao kexinzhao merged commit d3b8bff into PaddlePaddle:develop Oct 12, 2017
@kexinzhao kexinzhao deleted the decayed_adagrad_op branch October 12, 2017 06:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants