Offline Reinforcement Learning on Atari Breakout 1% DQN-replay v0
267.5Gamer-normalized ScoreDecision Transformer
Evaluation Results
| Method | Links | |
|---|---|---|
| Decision Transformerdataset variant=1% DQN-replay2021.06 | 267.5 | |
| CQLdataset variant=1% DQN-replay2021.06 | 211.1 | |
| Behavior Cloningdataset variant=1% DQN-replay2021.06 | 138.9 | |
| QR-DQNdataset variant=1% DQN-replay2021.06 | 17.1 | |
| REMdataset variant=1% DQN-replay2021.06 | 8.9 |