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

Support mixed precision training to speedup training #106

Closed
r9y9 opened this issue May 21, 2022 · 3 comments
Closed

Support mixed precision training to speedup training #106

r9y9 opened this issue May 21, 2022 · 3 comments
Assignees
@r9y9
Labels
enhancement New feature or request

Comments

@r9y9
Copy link
Collaborator

r9y9 commented May 21, 2022

https://pytorch.org/docs/stable/notes/amp_examples.html
@r9y9 r9y9 added the enhancement New feature or request label May 21, 2022
@r9y9 r9y9 self-assigned this May 21, 2022
r9y9 added a commit that referenced this issue Jun 3, 2022
@r9y9
implement mixed precision training for train/train_resf0
5f0ea38 
ref #106
r9y9 added a commit that referenced this issue Jun 4, 2022
@r9y9
mixed precision training support for cyclegan postfilter
7b42d59 
ref #106
@r9y9
Copy link
Collaborator Author

r9y9 commented Jun 4, 2022 •
edited

Done on the dev branch. GPU memory usage will be greatly reduced for training big models (only if the mixed precision training is supported by GPUs/CUDA). We'll see speedup too.

@r9y9
Copy link
Collaborator Author

r9y9 commented Jun 4, 2022

Double-checked training went well as usual.

@r9y9 r9y9 mentioned this issue Jun 5, 2022
Merged
@r9y9
Copy link
Collaborator Author

r9y9 commented Jun 5, 2022

Fixed by #109

@r9y9 r9y9 closed this as completed Jun 5, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@r9y9 r9y9

Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@r9y9