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

Feature/init grad in time #17653

Merged
merged 3 commits into from
May 29, 2019

Conversation

JiabinYang
Copy link
Contributor

@JiabinYang JiabinYang commented May 26, 2019

This PR using lazy Init Grad to save GPU memory in dygraph mode.

Make test on Resnet, and with Tesla K40m (GPU memory of 11439MiB) :
With lazy mode: max batch size is 56
With previous mode: max batch size is 29

@JiabinYang
Copy link
Contributor Author

No description provided.

@JiabinYang JiabinYang closed this May 26, 2019
@JiabinYang JiabinYang reopened this May 26, 2019
Copy link
Collaborator

@sneaxiy sneaxiy left a comment

Choose a reason for hiding this comment

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

LGTM.

Copy link
Collaborator

@sneaxiy sneaxiy left a comment

Choose a reason for hiding this comment

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

LGTM.

@JiabinYang JiabinYang merged commit effc555 into PaddlePaddle:develop May 29, 2019
@JiabinYang JiabinYang deleted the feature/init_grad_in_time branch January 18, 2021 08:36
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.

2 participants