Goal-conditioned Reinforcement Learning on Key-to-Door (10K Random Trajectories)
95.1Success RatePercentile Behavior Cloning
Evaluation results
| Method | Links | |
|---|---|---|
| Percentile Behavior Cloning2021.06 | 95.1 | |
| Decision Transformer2021.06 | 94.6 | |
| CQL2021.06 | 13.3 | |
| Random2021.06 | 3.1 | |
| BC2021.06 | 1.6 |