Offline Goal-Conditioned Reinforcement Learning on cube-triple 100M
Evaluation results
| Method | Links | |
|---|
| DQCMethod=Decoupled Q-chunking2025.12 | | 98 |
|---|
| NSBackup=n-step return backup2025.12 | | 93 |
|---|
| SHARSA2025.12 | | 83 |
|---|
| IQL2025.12 | | 66 |
|---|
| HFBC2025.12 | | 56 |
|---|
| FBC2025.12 | | 54 |
|---|
| OSBackup=1-step TD-backup2025.12 | | 47 |
|---|
| HIQL2025.12 | | 35 |
|---|
| DQC-naïveExecution=partial action chunk2025.12 | | 27 |
|---|
| QCStrategy=Q-chunking2025.12 | | 20 |
|---|