← Back to Test-Time Training
online logit bias
Test-Time TrainingUsed in
2 PRs
Best BPB
1.1248
Avg BPB
1.1429
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 218 | {"learning_rate":0.1,"momentum":0.9} |
| 330 | {"learning_rate":0.1,"enabled":false} |