← Back to Architecture
residual mixing
ArchitectureUsed in
6 PRs
Best BPB
1.0577
Avg BPB
1.2189
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 103 | — |
| 443 | — |
| 460 | — |
| 790 | {"layers":11,"dimensions":512,"mlp_multiplier":3.5,"mha":"8/8","bigramhash":8192,"xsa":"all layers"} |
| 1180 | — |
| 1527 | — |