← Back to Architecture

XSA4

Architecture
Used in
10 PRs
Best BPB
0.2841
Avg BPB
1.0147

Hyperparameters Across PRs

pr_numberparameters
416{"layers":11}
505{"layers":4}
531{"layers":4,"active_layers":[7,8,9,10]}
535{"layers":4}
610{"layers":4}
657{"heads":4}
672
685
807
864