Table 2 List of hyperparameters.

From: Deep hierarchical embedding for simultaneous modeling of GPCR proteins in a unified metric space

Hyperparameter Value
Kernel sizes 8, 12, 16, 20, 24, 28, 32, 36
Number of filters 256 \(\times \) 8
Dimension of embedding vectors 30
Number of hidden units in classifier 15
Learning rate 0.001
L2-regularizer on classifier 0.0005
Weight on center loss (\(\lambda _{C}\)) 0.01 0.3 0.5
Weight for center loss on family level (\(\omega _{C_{cls}}\)) 0.8 0.1 0.1
Weight for center loss on subfamily level (\(\omega _{C_{fam}}\)) 0.15 0.8 0.15
Weight for center loss on sub-subfamily level (\(\omega _{C_{sub}}\)) 0.05 0.1 0.75
Weight for cross-entropy loss with a softmax function on family level (\(\omega _{S_{cls}}\)) 0.8 0.1 0.1
Weight for cross-entropy loss with a softmax function on subfamily level (\(\omega _{S_{fam}}\)) 0.15 0.8 0.25
Weight for cross-entropy loss with a softmax function on sub-subfamily level (\(\omega _{S_{sub}}\)) 0.05 0.1 0.65