Implementing exponential moving average Lightning callback #257
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Lee, Kin Long Kelvin <[email protected]>
Signed-off-by: Lee, Kin Long Kelvin <[email protected]>
Signed-off-by: Lee, Kin Long Kelvin <[email protected]>
Signed-off-by: Lee, Kin Long Kelvin <[email protected]>
laserkelvin
added
enhancement
|
The SAM callback tests are failing - don't think it's due to things I've changed... |
melo-gonzo
approved these changes
Jul 19, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds an exponential moving average callback for improved generalization in training dynamics, in a modular fashion.
The callback works entirely independently of a given task: the EMA model updates are performed by this callback's training hooks, rather than by the task itself. That said, base methods in tasks were updated such that inference is done with the EMA model weights if they are available.
An example script is also provided showing how to use both SWA and EMA callbacks in a training run.