Reinforcement Learning on MuJoCo FetchReach v4
12.8Success RateLLM Iterative
Evaluation results
| Method | Links | |
|---|---|---|
| LLM IterativeTraining Episodes=3,0002026.05 | 12.8 | |
| Hand-CraftedTraining Episodes=3,0002026.05 | 8 | |
| LLM One-ShotTraining Episodes=3,0002026.05 | 7.5 | |
| No ShapingTraining Episodes=3,0002026.05 | 6.5 | |
| RNDTraining Episodes=3,0002026.05 | 6.5 |