← Back to Architecture
Hybrid GDN/Transformer
ArchitectureUsed in
1 PRs
Best BPB
1.2093
Avg BPB
1.2093
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 651 | {"layers":8,"ratio":"3:1"} |
| pr_number | parameters |
|---|---|
| 651 | {"layers":8,"ratio":"3:1"} |