PyTorch 1.x Reinforcement Learning Cookbook

How to do it...

We develop deep Q-learning using DQN as follows:

1. Import all the necessary packages:

>>> import gym>>> import torch>>> from torch.autograd import Variable>>> import random

The variable wraps a tensor and supports backpropagation.

2. Let's start with the __init__ method of the DQN class:

>>> class DQN(): ...     def __init__(self, n_state, n_action, n_hidden=50,                      lr=0.05): ...         self.criterion = torch.nn.MSELoss() ...         self.model = torch.nn.Sequential( ...                         torch.nn.Linear(n_state, n_hidden), ...                         torch.nn.ReLU(), ...                         torch.nn.Linear(n_hidden, n_action) ...                 ) ...         self.optimizer = torch.optim.Adam(                             self.model.parameters(), lr)

3. We now develop the training method, which updates the neural network with a data point:

>>> def update(self, ...