← Back to Architecture
layerwise LN scale
ArchitectureUsed in
3 PRs
Best BPB
1.0752
Avg BPB
1.1174
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 492 | — |
| 585 | — |
| 1584 | — |
| pr_number | parameters |
|---|---|
| 492 | — |
| 585 | — |
| 1584 | — |