Offlinerl

Repo for offline reinforcement learning methods

Included Dataset

Current support dataset:

• D4RL(website)

Planning to support:

• DQN Replay Dataset(website)

Methods

Behavior Clone methods:

• Behavior cloning

Basic Off-policy methods

• DDPG
• SAC
  • Basic SAC
  • Twin SAC without V

Offline RL methods:

• BCQ
  • Continuous BCQ