Human-Level Control Through Deep Reinforcement Learning

Source: https://storage.googleapis.com/deepmind-media/dqn/DQNNaturePaper.pdf
Authors: Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A. Rusu, Joel Veness, Marc G. Bellemare, Alex Graves, Martin Riedmiller, Andreas K. Fidjeland, Georg Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen King, Dharshan Kumaran, Daan Wierstra, Shane Legg & Demis Hassabis

This is a pytorch implementation of Deep Q Networks as described in the paper above.

Requirements + Versions

• Python 3.6
• PyTorch 1.0
• Pillow 5.3
• Open AI Gym

Running

$ python3 src/main.py