MetaGym

MetaGym provides abundant environments for benchmarking Reinforcement Learning and Meta Reinforcement Learning

Environments Updating

• LiftSim：Simulator for Evelvator Dispatching (Sep, 2019)

• Quadrotor: 3D Quadrotor simulator for different tasks (Mar, 2020)

• Quadrupedal: Quadrupedal robot adapting to different terrains (Seq, 2021)

• MetaMaze: 2D/3D maze generators for task generalization (Oct, 2021)

• MetaLocomotion: Locomotion simulator with diverse geometries (June, 2022)

• MetaLM: Meta language model dataset (Dec, 2022)

• Bandits: Bandits task generalization (Dec, 2022)