README.md

CS 294: Deep Reinforcement Learning, Fall 2017 - Homework

The course website : http://rail.eecs.berkeley.edu/deeprlcourse/

Here are my solutions for CS 294 in Pytorch. I covered this course remotely and therefore I don't guarantee my solutions are perfect. Some given codes originally depended on TensorFlow are modified for PyTorch. I worked the assignments to improve my Pytorch skills and understanding RL.

Dependencies:

Tensorflow is not required at all as I modifed all codes that depends on Tensorflow.

• PyTorch 0.4.0
• MuJoCo version 1.31
• OpenAI Gym 0.9.1

HW1

• Behavior Learning
• DAgger

HW2

• Policy Gradient

HW3

• Deep Q-Network(DQN) with target network on Atari