reinforcement-learning-awesome-varieties

• I have tried reinforcement learning basic approach to some environment sets.
• Also I tried PTAN, RAY a bit.
• Codes and scripts are referring to multiple textbooks, but with some original arrangements.
• Some codes includes playing learned model.

Approaches and Enviroments covered are: (just example)

• Actor-Critic, replay buffer
  • cartpole-v0 (gym)
• A2C (Advantage Actor-Critic)
  • Catcher-v0
  • HalfCheetahBulletEnv-v0
  • MinitaurBulletEnv-v0
  • PongNoFrameskip-v4
• A3C (Asynchronous Advantage Actor-Critic)
  • PongNoFrameskip-v4
• ACKTR (Actor-Critic using Kronecker-Factored Trust Region)
  • HalfCheetahBulletEnv-v0
• simple neural net (value function agent, policy gradient agent)
  • cartpole-v0
• D4PG (Distributed Distributional Deterministic Policy Gradients)
  • ant
  • HalfCheetahBulletEnv-v0
  • MinitaurBulletEnv-v0
• Dagger
  • FrozenLakeEasy-v0 (gym)
• DDPG (Deep Deterministic Policy Gradients)
  • MinitaurBulletEnv-v0
• Deep-Q
  • catcher
• Distributional DQN
  • PongNoFrameskip-v4
• Dueling DQN
  • PongNoFrameskip-v4
• Evolution
  • Catcher-v0
• GAN
  • Atari
• Noisy Networks
  • PongNoFrameskip-v4
• PPO (Proximal Policy Optimization)
  • cartpole-v0
  • HalfCheetahBulletEnv-v0
• Priority Replay Buffer
  • PongNoFrameskip-v4
• SAC (Soft Actor-Critic)
  • HalfCheetahBulletEnv-v0
• SARSA (state-Action-Reward-State-Action)
  • FrozenLakeEasy-v0
• TRPO (Trust Region Policy Optimization)
  • HalfCheetahBulletEnv-v0