← Back to Architecture

MLP 3.5x with LeakyReLU(0.5)^2

Architecture
Used in
1 PRs
Best BPB
1.1330
Avg BPB
1.1330

Submissions

PR #635by aryanbhosale
1.1330

Hyperparameters Across PRs

pr_numberparameters
635{"expansion_factor":3.5,"activation":"LeakyReLU(0.5)^2","hidden_dim":1792}