Python Deep Learning Projects

Module 4 – train_dqn.py

In this module, we include the memory() and  train() functions and also the calls to train and test the reinforcement learning model:

"""This module is used to train and test the DQN agent."""import randomimport numpy as npfrom agent_replay_dqn import agent, agent_action, replay, performance_plotfrom hyperparameters_dqn import *from test_dqn import testfrom keras import backend as kfrom collections import dequeimport gymenv = gym.make('CartPole-v1')# Set seed for reproducibilityseed_val = 456np.random.seed(seed_val)env.seed(seed_val)random.seed(seed_val)states = env.observation_space.shape[0]actions = env.action_space.ntraining_data = deque(maxlen=deque_len)def memory(state, new_state, reward, done, action): """Function ...