Table 1 List of Optimal Hyper-parameters value of proposed DNN model
From: Deep-ProBind: binding protein prediction with transformer-based deep learning model
List of parameters | Optimal values |
|---|---|
Activation functions | Tanh and SoftMax |
Regularization l2 | 0.001 |
Number of hidden layers | 4 |
Learning rates | 0.1 |
Number of Neurons at hidden layers | 85-43-16-4 |
Optimizer | SGD method |
Updater | ADAGRAD function |
Weight initialization function | XAVIER function |
Seed | 12345L |
Training Epoch | 50 |
Dropout | 0.25 |
Momentum | 0.9 |