layout: "contents" title: Basic Usage
There is a UI application which allows you to manually control the agent with the arrow keys:
./minigrid/manual_control.py
The environment being run can be selected with the --env
option, eg:
./minigrid/manual_control.py --env MiniGrid-Empty-8x8-v0
If you want to train an agent with reinforcement learning, I recommend using the code found in the torch-rl repository. This code has been tested and is known to work with this environment. The default hyper-parameters are also known to converge.
A sample training command is:
cd torch-rl
python3 -m scripts.train --env MiniGrid-Empty-8x8-v0 --algo ppo