[RLlib] Example script: Simple league-based self-play w/ the open spiel markov soccer env. #17077
Conversation
This file contains 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
…cy_map_as_lru_cache # Conflicts: # rllib/policy/tf_policy.py
…nto policy_map_as_lru_cache
…cy_map_as_lru_cache
…cy_map_as_lru_cache # Conflicts: # rllib/evaluation/rollout_worker.py
…cy_map_as_lru_cache
…eague_based_example
…cy_map_as_lru_cache_league_based_example
…cy_map_as_lru_cache_league_based_example
…cy_map_as_lru_cache_league_based_example
sven1977
added
the
tests-ok
sven1977
removed
the
tests-ok
michaelzhiluo
approved these changes
Jul 21, 2021
…cy_map_as_lru_cache_league_based_example
7 tasks
7 tasks
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Outdated suggestions cannot be applied.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.
Example script: Simple league-based self-play w/ the open spiel a) markov soccer and b) connect-4 envs.
The script shows how to set up a simple league consisting of 3 types of policies, similar to DeepMind's StarCraft II setup:
Starting with a single "main" and two randomly initialized league- & main-exploiters, the league is further augmented during training. This is achieved by measuring each policy's win-rate and then cloning a policy when it has reached a certain win-rate. The new clone is either frozen or keeps being trained (randomly chose between these two options).
In the end, we'll have a main policy that robustly plays against different opponent strategies and against catastrophic forgetting.
Why are these changes needed?
Related issue number
Checks
scripts/format.shto lint the changes in this PR.