Reinforcement Learning on Atari 2600 (10M Frames, 6 Game Subset)
2,950.1Alien ScoreDreamer-v2
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Dreamer-v2Training frames=10M, Evaluation criteria=Best of 3 runs2021.11 | 2,950.1 | 3,100.8 | 57 | -13.6 | 16,002.8 | 13,552.9 | |
| MBEC++Training frames=10M, Evaluation criteria=Best of 3 runs2021.11 | 1,991.2 | 7,724 | 372.2 | 17.6 | 11,386.4 | 12,148.5 |