create_model: Initialize a fully-connected multi-class neural network

In c7rishi/hidgenclassifier: Functions for Bayesian hierarchical hidden genome classifier

Initialize a fully-connected multi-class neural network

Description

Initialize a fully-connected multi-class neural network

Usage

create_model(
  learning_rate = 0.001,
  weight_decay = 0,
  dropout = 0,
  num_dense_layers = 1,
  num_dense_nodes = 5,
  activation = "relu",
  seed = NULL,
  input_shape,
  num_classes
)

Arguments

learning_rate	Learning rate for Adam optimizer
weight_decay	L2 regulariation paramter
dropout	Dropout rate
num_dense_layers	Number of hidden layers
num_dense_nodes	Number of nodes per hidden layer (all hidden layers have the same size)
activation	Name of activation function ("relu", "softmax", "softplus", "elu", "tanh", "sigmoid", etc)
seed	Seed for weight initialization and dropout
input_shape	Number of features
num_classes	Number of classes

Value

A compiled Keras model

Author(s)

Zoe Guan summary(model)

c7rishi/hidgenclassifier documentation built on June 14, 2024, 11:10 a.m.