← Back to Architecture
Token-routed MoE
ArchitectureUsed in
1 PRs
Best BPB
1.4072
Avg BPB
1.4072
Submissions
Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 377 | {"experts":4} |
| pr_number | parameters |
|---|---|
| 377 | {"experts":4} |