← Back to Architecture
attention scaling
ArchitectureUsed in
1 PRs
Best BPB
1.5990
Avg BPB
1.5990
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 2159 | {"num_heads":4} |
| pr_number | parameters |
|---|---|
| 2159 | {"num_heads":4} |