DeepMind Control
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Reinforcement Learning | DeepMind Control Cartpole Balance Sparse | Steps to 75% Return150,700 | 11 | |
| Behavior Cloning | DeepMind Control (DMC) suite seen/unseen embodiments | Hopper Hop Score49.1 | 9 | |
| Reinforcement Learning | DeepMind Control Reacher Hard | Steps to 75% Return (k)589.7 | 8 | |
| Vision-based robot policy learning | DeepMind Control | Stand Performance89.1 | 8 | |
| Reinforcement Learning | DeepMind Control Walker Walk | Steps to 75% Return116.8 | 6 | |
| Reinforcement Learning | DeepMind Control Walker Stand | Steps to 75% Return120 | 6 | |
| Reinforcement Learning | DeepMind Control Reacher Easy | Steps to 75% Return Threshold205,000 | 6 | |
| Reinforcement Learning | DeepMind Control Quadruped Walk | Steps to Reach 75% Return330,500 | 6 | |
| Reinforcement Learning | DeepMind Control Pendulum Swingup | Steps to 75% Return (k)76,600 | 6 | |
| Reinforcement Learning | DeepMind Control Hopper Hop | Steps to 75% Return153,500 | 6 | |
| Reinforcement Learning | DeepMind Control Finger Turn Hard | Steps to 75% Return333.4 | 6 | |
| Reinforcement Learning | DeepMind Control Finger Turn Easy | Steps to 75% Return77,300 | 6 | |
| Reinforcement Learning | DeepMind Control Finger Spin | Steps to 75% Return89,100 | 6 | |
| Reinforcement Learning | DeepMind Control Cup Catch | Steps to 75% Return104,700 | 6 | |
| Reinforcement Learning | DeepMind Control Cartpole Swingup | Steps to 75% Return278,800 | 6 | |
| Reinforcement Learning | DeepMind Control Cartpole Balance | Steps to 75% Return203,100 | 6 | |
| Reinforcement Learning | DeepMind Control Acrobot Swingup | Steps to 75% Return475,000 | 6 | |
| Reinforcement Learning | DeepMind Control Walker Run | Steps to 75% Return205,000 | 5 | |
| Reinforcement Learning | DeepMind Control Hopper Stand | Steps to 75% Return (k)245.2 | 5 | |
| Reinforcement Learning | DeepMind Control Cheetah Run | Environment Steps to 75% Return (k)145,400 | 5 | |
| Reinforcement Learning | DeepMind Control Cartpole Swingup Sparse | Environment Steps to 75% Return383,100 | 5 | |
| Continuous Control | DeepMind Control 3M steps | Acrobot Swingup Score422 | 5 | |
| Continuous Control | DeepMind Control 1.5M steps | Acrobot Swingup Score272 | 5 | |
| Reinforcement Learning | DeepMind Control Quadruped Run | Steps to 75% Return345,000 | 4 | |
| Continuous Control | DeepMind Control | Walker-stand Success Rate89.1 | 4 |
Showing 25 of 30 rows