Reinforcement Learning Reimplementations

Reimplementations of various reinforcement learning algorithms:

• Actor-critic (including policy gradients)
• Value-based (Q-learning)
• Unsupervised (reward-free i.e. curiosity)