[BugFix] Account for terminating data in SAC losses #2606
Merged
+162
−8
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/2606
Note: Links to docs will display an error until the docs builds have been completed. ❌ 17 New Failures, 1 Unrelated FailureAs of commit 18bf7f5 with merge base d90b9e3 ( NEW FAILURES - The following jobs have failed:
BROKEN TRUNK - The following job failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
vmoens
added a commit
that referenced
this pull request
Nov 25, 2024
ghstack-source-id: dc1870292786c262b4ab6a221b3afb551e0efb9b Pull Request resolved: #2606
facebook-github-bot
added
the
CLA Signed
vmoens
added a commit
that referenced
this pull request
Nov 25, 2024
ghstack-source-id: dc1870292786c262b4ab6a221b3afb551e0efb9b Pull Request resolved: #2606
Closed
3 tasks
| # Check done state and avoid passing these to the actor | ||
| done = next_tensordict.get(self.tensor_keys.done) | ||
| if done is not None and done.any(): | ||
| next_tensordict_select = next_tensordict[~done.squeeze(-1)] |
There was a problem hiding this comment.
The done shape could be more extended than the batch shape, this line is breaking in multiagent settings
There was a problem hiding this comment.
Then we need a test that covers this use case!
Can you draft one for me?
There was a problem hiding this comment.
The SOTA ci picked up on this. Both SAC scripts are failing
There was a problem hiding this comment.
Ah i didn't see (SOTA is broken bc of dreamer so I didn't check)
we should have tests that are not in SOTA, SOTA is there to test that scripts run smoothly, not features. The scripts are not part of the core lib - we can arbitrarily decide to ditch them, the rest of the lib should still work.
There was a problem hiding this comment.
Yeah I long wanted to make some tests for multiagent data in losses, will get to it when I have time.
Right now just crunching on writing thesis and satisfying BenchMARL users in free time.
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.
Stack from ghstack (oldest at bottom):