[RLlib] Introduce IMPALA off_policyness test with GPU #31485
Merged
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: Artur Niederfahrenhorst <artur@anyscale.com>
Signed-off-by: Artur Niederfahrenhorst <artur@anyscale.com>
gjoliver
reviewed
Jan 9, 2023
| @@ -71,7 +71,7 @@ def __init__( | |||
| """ | |||
| self.capacity = capacity | |||
| self.replay_ratio = replay_ratio | |||
| self.replay_proportion = None | |||
| self.replay_proportion = 1 | |||
There was a problem hiding this comment.
why do you flip this flag here? it looks important.
There was a problem hiding this comment.
Uh good point. That's by accident. I wondered why it was None.
I'll change it back and ping you after tests are green again.
Signed-off-by: Artur Niederfahrenhorst <attaismyname@googlemail.com>
AmeerHajAli
pushed a commit
that referenced
this pull request
Jan 12, 2023
Signed-off-by: Artur Niederfahrenhorst <artur@anyscale.com>
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Outdated suggestions cannot be applied.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.
Signed-off-by: Artur Niederfahrenhorst artur@anyscale.com
Why are these changes needed?
IMPALA runs with a GPU learner "by original design". A compilation test without GPU is fine but something like off_policyness depends on learning/sampling speed and we should test it for our standard impala settings (gpu=1, num_rollout_workers=2) - with a GPU and without competing over resources. Only then will we be able to get meaningful measurements of it's off-policyness.
This PR also tries to deflake IMPALA tests.