D4RL / Adroit legacy benchmarks
Farama-Foundation
Historical offline reinforcement-learning benchmarks, including Adroit dexterous-hand tasks.
What you can explore
Understand older dexterous offline-learning results and locate the successor environment and dataset projects.
Before you use it
The maintainers moved environments to Gymnasium projects and offline datasets to Minari. Prefer the maintained successors for new work; old dependencies and evaluation versions differ.
Original contributors
D4RL authors and Farama Foundation contributors
License / access: Code: Apache 2.0 · datasets: CC BY 4.0 unless otherwise noted
Keep exploring
Minari
Tools for storing, sharing and using offline reinforcement-learning datasets.
Gymnasium Robotics
MuJoCo robotics environments with Gymnasium interfaces, including dexterous hand tasks.
DAPG
Demonstration-augmented policy-gradient learning for dexterous hand manipulation.
Editorial reference · checked 2026-09-22