Hands-On Reinforcement Learning for Games

Training in Habitat

At the time of writing, Habitat was quite new but showed amazing potential, especially for training agents. This means the environment currently only has a simple and PPO agent implementation in which you can quickly train agents. Of course, since Habitat uses PyTorch, you could probably implement one of the other algorithms we have covered. In the next exercise, we finish off by looking at the PPO implementation in Habitat and how to run it:

1. Download and install the Habitat API with the following commands:

git clone --branch stable [e-mail]:facebookresearch/habitat-api.git cd habitat-api pip install -r requirements.txt python setup.py develop --all

2. At this point, you can use the API in a number of ways. We will ...