- 💪 Challenge Page: https://www.aicrowd.com/challenges/dr-derk-s-mutant-battlegrounds
- 🏆 Leaderboard: https://www.aicrowd.com/challenges/dr-derk-s-mutant-battlegrounds/leaderboards
This starter kit contains an agent trained to play Dr. Derk's Mutant Battlegrounds.
pip install -U gym-derk
For more information, refer to the official documentation.
You can start with the default bot.py or create your own agent that file. You can also your own script containing DerkPlayer class format present in bot.py.
The functions that are required in the DerkPlayer class are:
__init__
: To initialize your player and will have the parametersn_agents
andaction_space
signal_env_reset
: To signal to your player that the environment has been reset and has the parametersobservation
take_action
: This is the function where the observation will be passed and the actions for each agent are returned.
Please don't change the definition of these functions. You can add more if you want to.
Run the python script:
python run.py -n 1
The python script has the following command line arguments:
usage: run.py [-h] [-p1 FILE_NAME_OF_PLAYER1] [-p2 FILE_NAME_OF_PLAYER2]
[-n NUMBER_OF_ARENAS] [--fast]
optional arguments:
-h: to display help messages
-p1: The name of the file that contains player 1
-p2: The name of the file that contains player 2
-n: The number of arenas to run parallely (defaults to 2)
--fast: To enable turbo mode or not (cuts down episode time by half by speeding up animations)
For example, if you have bot.py
and oldbot.py
inside the agent
directory, you can have a fight between them by running
python run.py -p1 bot -p2 oldbot