The exercises in this section are intended to introduce you to Unity ML-Agents in more detail. If your preference is not to use ML-Agents as a training framework, then move on to the next section and the end of this chapter. For those of you still here, ML-Agents on its own is a powerful toolkit for quickly exploring DRL agents. The toolkit hides most of the details of DRL but that should not be a problem for you to figure out by now:
- Set up and run one of the Unity ML-Agents sample environments in the editor to train an agent. This will require that you consult the Unity ML-Agents documentation.
- Tune the hyperparameters of a sample Unity environment.
- Start TensorBoard and run it so that it collects logs from the Unity runs folder. This will allow you to watch the training performance of the agents being trained with ML-Agents.
- Build a Unity environment and train...
