[BugFix] ActionMask is compatible with composite action specs #3022
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3022
Note: Links to docs will display an error until the docs builds have been completed. ❌ 5 New Failures, 21 Pending, 22 Unrelated FailuresAs of commit 32c62e4 with merge base 5a13341 ( NEW FAILURES - The following jobs have failed:
FLAKY - The following jobs failed but were likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
louisfaury
commented
Jun 20, 2025
Comment on lines
-8776
to
-8777
| >>> env = TransformedEnv(base_env, ActionMask()) | ||
| >>> r = env.rollout(10) |
There was a problem hiding this comment.
Already in the lines above.
Comment on lines
+8729
to
+8732
| This transform is useful to ensure that randomly generated actions | ||
| respect legal actions, by masking the action specs. | ||
| It reads the mask from the input tensordict after the step is executed, | ||
| and adapts the mask of the finite action spec. |
There was a problem hiding this comment.
Took the occasion to refine a bit the docstring.
Comment on lines
-8814
to
-8817
| action_spec = self.container.full_action_spec | ||
| keys = self.container.action_keys | ||
| if len(keys) == 1: | ||
| action_spec = action_spec[keys[0]] |
There was a problem hiding this comment.
tbh I could not really see why this check was in place. I hope I did not miss any chesterton fence!
There was a problem hiding this comment.
It looks good, there is some confusion in the existing code about in_keys and action_keys but this seems to be aligned with what we want.
We should maybe have a validation step when we register a transform in an env. For example here, when we set the container to the transform, we should check that the action_key arg is part of the action_keys of the parent env.
Comment on lines
-8842
to
-8851
| action_spec = self.action_spec | ||
| mask = tensordict.get(self.in_keys[1], None) | ||
| if mask is not None: | ||
| mask = mask.to(action_spec.device) | ||
| action_spec.update_mask(mask) | ||
|
|
||
| # TODO: Check that this makes sense | ||
| with _set_missing_tolerance(self, True): | ||
| tensordict_reset = self._call(tensordict_reset) | ||
| return tensordict_reset |
There was a problem hiding this comment.
I think this entire logic is already handled by _call
vmoens
added
bug
vmoens
force-pushed
the
lf/action-mask-many-actions
branch
from
630b39d to
32c62e4
Compare
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.
Description
ActionMaskwas not compatible with composite action specs because of the lines:that afaiu do not serve a real purpose.
This PR removes this check, and allows tu use
ActionMaskto mask one factor of a composite action spec.Types of changes
What types of changes does your code introduce? Remove all that do not apply:
Checklist
Go over all the following points, and put an
xin all the boxes that apply.If you are unsure about any of these, don't hesitate to ask. We are here to help!