Deep RL elements

Deep RL algorithm with pytorch (written in jupyter notebook). Algorithms are tested on Atari Games, mostly with Pong. Feel free to ask questions and report bugs with Issues. :)

1. Deep Q-learning

2. DQN with prioritized experience replay (PER)

3. Double DQN (DDQN)

4. Dueling DQN

5. DQN with soft update (Soft DQN)

6. DQN with noisy net (Noisy DQN)

7. DQN with intrinsic curiosity module (ICM DQN)

8. Deep recurrent Q-learning (DRQN)

9. N-step DQN

Prerequisite

• torch 1.4.0
• gym[atari]
• tensorboardX

Other useful deep RL algorithm zoos (pytorch)

• Reinforcement-Learning
• DeepRL-Tutorials
• Deep-Reinforcement-Learning-Algorithms-with-PyTorch