A set of offline RL algorithm with tensorflow-keras.
- Use D4RL dataset
- For the purpose of result reproducibility.
- Heavily borrow most test concepts from TD3+BC.
A Minimalist Approach to Offline RL, Fujumoto et al [paper], [repo]
Offline Reinforcement Learning with Implicit Q-Learning, Kostrikov et al [paper][repo(jax)][repo(pytorch)]