Merlin

The dataset used in the paper is a collection of trajectories for goal-conditioned reinforcement learning tasks.

BibTex: