DeepMind Control Suite

DeepMind's standardized MuJoCo-based continuous-control RL benchmark suite, the canonical evaluation for policy-gradient and model-based RL.

Open

Publisher: Google DeepMind
Year: 2018
Venue: preprint
ArXiv: arxiv.org/abs/1801.00690
URL: arxiv.org/abs/1801.00690
Code: github.com/google-deepmind/dm_control
Authors: 12
Hosting: External sourcelicense unknown

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text: arxiv.org/abs/1801.00690
TL;DR: semanticscholar.org/paper/a9a3ed69c94a3e1c08ef1f833d9199f57736238b
Code: github.com/google-deepmind/dm_control

Attribution policy →

Introduces 3 artifacts - 1 eval, 2 tools

TL;DR

Semantic Scholar

The DeepMind Control Suite is a set of continuous control tasks with a standardised structure and interpretable rewards, intended to serve as performance benchmarks for reinforcement learning agents.

Artifacts

Evals

DeepMind Control Suite

Tools

DeepMind Control Suite DM Control RL Env (Hugging Face)

Authors

Abbas Abdolmaleki Alistair Muldal Andrew Lefrancq David Budden Diego de Las Casas Josh Merel Martin Riedmiller Timothy Lillicrap Tom Erez Yazhe Li Yotam Doron Yuval Tassa