Prevent NaN reward propagation in Humanoid tasks - #353
Open
sylvesterkaczmarek wants to merge 2 commits into
Open
sylvesterkaczmarek wants to merge 2 commits into
sylvesterkaczmarek wants to merge 2 commits into
Conversation
sylvesterkaczmarek
force-pushed
the
fix/humanoid-nan-reward
branch
from
September 10, 2026 15:58
36cf4bb to
8dadf4c
Compare
btaba
approved these changes
Sep 10, 2026
This branch has not been deployed
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Prevent NaN physics states from returning a NaN learner-facing reward in Humanoid tasks. Keep the existing termination condition and use
jp.whereto select a zero reward for invalid states.Rebased onto upstream
mainat4057c147714b6ac09b395377f1a1724bbeacc4d3, preserving both newer Warp overflow-warning tests. The regression test now covers NaNs in eitherqposorqvel.Validation
JAX_PLATFORMS=cpu python -m pytest -q mujoco_playground/_src/dm_control_suite/dm_control_suite_test.py -k 'Humanoid or humanoid'5 passed: two invalid-state regressions and the existing Humanoid stand, walk, and run smoke tests. Both regressions fail on unpatched
mainwithnan != 0.0.git diff --checkpasses. Ruff reports the same seven pre-existing findings in the two changed files, with no additional findings. The full suite and GPU execution were not run.Fixes #240.