yhyu13/C51-DDPG

This is a TensorFlow implementation of DeepMind's A Distributional Perspective on Reinforcement Learning.(C51-DDPG)

Standard implementation of the DDPG algorithm https://arxiv.org/abs/1509.02971 Coupled with DeepMind's newest ditributional bellman equation update, chekout critic_network.py (loss function) and ddpg.py (train function) for details. https://arxiv.org/pdf/1707.06887.pdf FILES: actor_network.py: The code for structure of the actor network
critic_network.py: The code for structure of the critic network
ou_noise.py:The random noise generator
ddpg.py:The code for the ddpg algorithm
gym_ddpg.py:Running ddpg on the environment