← Back to Architecture
XSA-all
ArchitectureUsed in
4 PRs
Best BPB
0.5440
Avg BPB
0.9160
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 569 | {"layers":11} |
| 756 | — |
| 795 | {"layers":11,"dim":512,"gqa_heads":"8/4"} |
| 825 | — |