This is the code for reproducing the results of the paper Offline RL with No OOD Actions: In-Sample Learning via Implicit Value Regularization accepted as Notable-top-5% at ICLR'2023.

The discrete version of IVR on Atari datasets can be found at https://github.com/ryanxhr/Discrete_IVR.

Our code is built on the jax version code of IQL (https://github.com/ikostrikov/implicit_q_learning). Paper reuslts can be reproduced by running ./run_mujoco.sh, ./run_antmaze.sh and ./run_kitchen.sh.

@inproceedings{xu2023offline,
  title  = {Offline RL with No OOD Actions: In-Sample Learning via Implicit Value Regularization},
  author = {Haoran Xu, Li Jiang, Jianxiong Li, Zhuoran Yang, Zhaoran Wang, Victor Wai Kin Chan, Xianyuan Zhan},
  year   = {2023},
  booktitle = {International Conference on Learning Representations},
}