DeepMind Control Suite

Introduced by Tassa et al. in DeepMind Control Suite

The DeepMind Control Suite (DMCS) is a set of simulated continuous control environments with a standardized structure and interpretable rewards. The tasks are written and powered by the MuJoCo physics engine, making them easy to identify. Control Suite tasks include Pendulum, Acrobot, Cart-pole, Cart-k-pole, Ball in cup, Point-mass, Reacher, Finger, Hooper, Fish, Cheetah, Walker, Manipulator, Manipulator extra, Stacker, Swimmer, Humanoid, Humanoid_CMU and LQR.

Source: Unsupervised Learning of Object Structure and Dynamics from Videos

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
Continuous Control	DeepMind Cheetah Run (Images)	DreamerV1
Continuous Control	DeepMind Cup Catch (Images)	DrQ
Continuous Control	DeepMind Walker Walk (Images)	DrQ
Continuous Control (100k environment steps)	DeepMind Finger Spin (Images)	PlayVirtual
Continuous Control (100k environment steps)	DeepMind Cartpole Swingup (Images)	PlayVirtual
Continuous Control (100k environment steps)	DeepMind Cheetah Run (Images)	PlayVirtual
Continuous Control (100k environment steps)	DeepMind Walker Walk (Images)	PlayVirtual
Continuous Control (500k environment steps)	DeepMind Finger Spin (Images)	PlayVirtual
Continuous Control (500k environment steps)	DeepMind Cartpole Swingup (Images)	PlayVirtual
Continuous Control (500k environment steps)	DeepMind Cheetah Run (Images)	PlayVirtual
Continuous Control (500k environment steps)	DeepMind Walker Walk (Images)	PlayVirtual