D4RL is a collection of environments for offline reinforcement learning. These environments include Maze2D, AntMaze, Adroit, Gym, Flow, FrankKitchen and CARLA.
433 PAPERS • 16 BENCHMARKS