-
Notifications
You must be signed in to change notification settings - Fork 5.8k
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.
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
[RLlib] Cleanup examples folder #16: Add missing docstrings to 2 connector example scripts. #45864
[RLlib] Cleanup examples folder #16: Add missing docstrings to 2 connector example scripts. #45864
Conversation
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
…nup_examples_folder_16_docstrings_for_connector_examples
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
# Define env-to-module-connector pipeline for the new stack. | ||
def _env_to_module_pipeline(env): | ||
return [ | ||
AddObservationsFromEpisodesToBatch(), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are the rest of the connectors now default?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
All off-the-shelf connectors now directly operate on the episodes now (except for framestacking for performance reasons):
- read last obs from episode
- alter that last obs
- directly write it back into the episode via
episode.set_observations
.
This way, RLlib's default pieces (which come after the user defined ones) only see the already altered episode and thus properly extract the observations from those and put them into the RLModule batch.
This way, things work more uniformly. Reading the last obs from an episode and writing it back is super fast, b/c we are still operating on simple lists in the EnvRunner (not numpy'ized yet).
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Signed-off-by: sven1977 <[email protected]>
Cleanup examples folder #16: Add missing docstrings to 2 connector example scripts.
Why are these changes needed?
Related issue number
Checks
git commit -s
) in this PR.scripts/format.sh
to lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/
under thecorresponding
.rst
file.