Allow for saving the PPOTrainer value model (critic model) #3308
Conversation
This file contains hidden or 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
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Outdated suggestions cannot be applied.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 functionality for
save_value_model: bool=Falsewhich determines whether to save the value model that is trained by PPOTrainer.During training, PPO puts quite a lot of effort training the value model to be good at predicting the reward the policy will receive. That sounds like a valuable tool, and it is a shame that previously we had simply been discarding it. This PR adds functionality for keeping the value model.
I requested this feature here, and this PR resolves that.
It also resolves the error described in this issue: no attribute 'policy' when pushing PPOTrainer to hub using example ppo.py script #3301
I'm unsure whether the way I fixed it, the early return on line 338 in PPOTrainer, is legal. Please check!
I would love your feedback, @qgallouedec!