[BugFix] Recurrent policy auto-register with policy_factory#3753
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3753
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 15 New Failures, 20 Unclassified FailuresAs of commit ae72264 with merge base 0a01ee8 ( NEW FAILURES - The following jobs have failed:
UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
This was referenced May 15, 2026
This was referenced May 17, 2026
vmoens
added a commit
that referenced
this pull request
May 18, 2026
Fixes two latent bugs the auto-register path exposed: - `_maybe_append_env_transforms_from_module` wrapped the new transforms in `Compose`, which parented the children to a throwaway container. `InitTracker` / `TensorDictPrimer` then saw `parent=None` and `env.action_spec` silently broke. Append each transform directly so their parent stays the live env. - `TensorDictPrimer.transform_input_spec` crashed on envs with no `full_state_spec`. Initialize an empty `Composite` first, mirroring `StepCounter`'s defensive default. Add a regression test pinning both: after auto-register the env exposes `is_init` and `recurrent_state`, `env.action_spec` stays reachable, and every transform has a live parent. ghstack-source-id: b6a6777 Pull-Request: #3753
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
1 participant
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):
Fixes two latent bugs the auto-register path exposed:
_maybe_append_env_transforms_from_modulewrapped the new transformsin
Compose, which parented the children to a throwaway container.InitTracker/TensorDictPrimerthen sawparent=Noneandenv.action_specsilently broke. Append each transform directly sotheir parent stays the live env.
TensorDictPrimer.transform_input_speccrashed on envs with nofull_state_spec. Initialize an emptyCompositefirst, mirroringStepCounter's defensive default.Add a regression test pinning both: after auto-register the env exposes
is_initandrecurrent_state,env.action_specstays reachable, andevery transform has a live parent.