[RLlib] [Connector] Fix EnvRunnerV2's handling of soft_horizon episodes. #30434
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Outdated suggestions cannot be applied.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: Jun Gong [email protected]
Why are these changes needed?
It is correct that for soft_reset, we don't reset agent_collector data.
But then, we need to make sure obs from these soft_reset steps gets added to to_eval for policy eval, so that env can continue to advance.
for soft_reset case, we shouldn't add an artificial obs-only, fake init_obs.
Related issue number
Checks
git commit -s
) in this PR.scripts/format.sh
to lint the changes in this PR.