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

fix(train): fix lr_decay not working #291

Merged

34j merged 1 commit into voicepaw:main from guranon:fix-lr-decay

Apr 11, 2023

Contributor

guranon commented Apr 11, 2023 •

edited by 34j

Loading

Since we're using manual optimization, we need to call step() on each of the learning rate schedulers.
This steps learning rates once per epoch so as to match the behavior before the lightning rewrite.
Relevant docs: https://lightning.ai/docs/pytorch/stable/common/optimization.html#learning-rate-scheduling

Closes #292


          fix(train): step lr schedulers at end of epoch

2e2c081

guranon force-pushed the fix-lr-decay branch from 1d05b8e to 2e2c081 Compare

April 11, 2023 05:31

Contributor Author

guranon commented Apr 11, 2023

Force push was to change the commit message (TIL about conventional commits)

34j merged commit 3af223e into voicepaw:main

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

Labels

None yet