← Back to Architecture

Value Residual (ResFormer)

Architecture
Used in
2 PRs
Best BPB
1.1330
Avg BPB
1.1379

Submissions

PR #516by Asukabot0
1.1428
PR #635by aryanbhosale
1.1330

Hyperparameters Across PRs

pr_numberparameters
516—
635—