Reinforcement Learning Policy Success on Full-Mix (test)
92.5Group 1 Success RatioDreamControl-v2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DreamControl-v2Train Data=Full-Mix, # Samples (K)=29.62026.03 | 92.5 | 98.5 | |
| DreamControl-v2Train Data=AMASS + Nymeria, # Samples (K)=26.42026.03 | 86 | 98.5 | |
| DreamControlTrain Data=Zero-Shot, # Samples (K)=-2026.03 | 4 | 99.5 | |
| DreamControl-v2Train Data=AMASS, # Samples (K)=21.82026.03 | 3 | 98.5 |