Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix envpool bug with missing mask key #3026

Merged
merged 1 commit into from
Aug 15, 2022
Merged

Fix envpool bug with missing mask key #3026

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 9 additions & 5 deletions gym/utils/step_api_compatibility.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,12 +56,15 @@ def step_to_new_api(
"TimeLimit.truncated" not in infos
or (
"TimeLimit.truncated" in infos
and not infos["_TimeLimit.truncated"][i]
and not infos["TimeLimit.truncated"][i]
)
) # vector env, dict info api, if mask is False, it's the same as TimeLimit.truncated attribute not being present for env 'i'
)
# vector env, dict info api, for env i, vector mask `_TimeLimit.truncated` is not considered, to be compatible with envpool
# For env i, `TimeLimit.truncated` not being present is treated same as being present and set to False.
# therefore, terminated=True, truncated=True simultaneously is not allowed while using compatibility functions
# with vector info
)
):

terminateds.append(dones[i])
truncateds.append(False)

Expand All @@ -80,10 +83,11 @@ def step_to_new_api(
truncateds.append(True)
else:
# This means info["TimeLimit.truncated"] exists but is False, which means the core environment had already terminated,
# but it also exceeded maximum timesteps at the same step.
# but it also exceeded maximum timesteps at the same step. However to be compatible with envpool, and to be backward compatible
# truncated is set to False here.
assert dones[i]
terminateds.append(True)
truncateds.append(True)
truncateds.append(False)

return (
observations,
Expand Down