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.

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

Dygraph Recompute: support amp #33251

Merged
merged 3 commits into from
Jun 3, 2021
Merged

Dygraph Recompute: support amp #33251

merged 3 commits into from
Jun 3, 2021

Conversation

JZ-LIANG
Copy link
Contributor

@JZ-LIANG JZ-LIANG commented Jun 1, 2021 •
edited
Loading

PR types

Function optimization

PR changes

OPs

Describe

Dygraph Recompute support amp
the usage of recompute please refer to PR

Tested on PaddleNLP Bert-base model and 32G V100 GPU, we have the following performance:

bsz MEM (MB) token/s avg loss @ 4w step
baseline 64 25248 49.30 3.4147
recompute 64 9466 37.50 3.4162
amp 64 16956 129.80 3.4239
amp-recompute 64 8782 100.29 3.4121

@paddle-bot-old
Copy link

paddle-bot-old bot commented Jun 1, 2021

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

ForFishes
ForFishes approved these changes Jun 3, 2021
View reviewed changes
Copy link
Member

@ForFishes ForFishes left a comment

Choose a reason for hiding this comment

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

LGTM

@ForFishes ForFishes merged commit 8752c91 into PaddlePaddle:develop Jun 3, 2021
@FesianXu
Copy link

请问2.1.3支持support amp的recompute吗

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

@ForFishes ForFishes ForFishes approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@JZ-LIANG @FesianXu @ForFishes