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Outdated suggestions cannot be applied.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.
Add the Chrome T-Rex rush game to PLE. It should be a fun game for learning & playing. 😄
Here's a random agent playing the game:
Game Spec
Observation Space
Action Space
The reason we add a
STAND
action is that somehowgym-ple
doesn't propogateNO_OP
to our game which makes our dino not able to stand up again after ducked once.Rewards
0 for ticking, 1 for passing each obstacle and -5 for game over.
We use predefined rewards from base game.
We also limit a high score to 50 (could be configured with max_score). A fun fact is that when I first try to train my agent, there's an episode lasts about 2 hours (passed 3K obstacles).
Simplification
We removed cloud and high score board in this implementation to make the observation cleaner for agents. Maybe we should add an option for this?
Credit
The implementation depends largely on @shivamshekhar 's Chrome-T-Rex-Rush.