← Back to Architecture

ValueEmbedding

Architecture
Used in
11 PRs
Best BPB
0.6364
Avg BPB
1.0726

Hyperparameters Across PRs

pr_numberparameters
768{"layers":[5,6,7,8,9,10]}
808{"layers":[9,10],"dimension":128}
849{"dim":128,"layers":[9,10]}
1066
1070{"dimensions":128}
1089{"layers":[9,10]}
1098
1117
1118
1126{"layers":[9,10]}
1169{"layers":[9,10]}