I've added a demo Frozen Lake simulation that uses basic Q-learning. It can be used to easily check that your Open Gymnasium environment is set-up correctly.
Firstly, make sure that pipenv
is installed using pip
. Then:
- Change into the
gym-frozen-lake
directory. - Run
pipenv sync
orpipenv install
to install the dependencies from thePipfile.lock
file. - You might need to run
pipenv shell
before this. In either case, run it now to activate the virtual environment. - You can then execute the Frozen Lake demo using
pipenv run python3 frozen_lake.py
. It should print the state space and action space, then the rewards from training. Finally it plays 3 games and prints the results of each. - Change the
render_mode
tohuman
to activate the simulation and see the training games being played "live".