model.py

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout

num_features = X_train.shape[1]

config = {
           'data': data,
           'train_test_ratio': 0.2,
           'model_config': {
               'model_name': 'house_pricing_model',
               'layers': {
                        'first_layer': 12, 
                        'second_layer': 5, 
                        'output_layer': 1
                       },
                'activations': ['relu', 'relu', None], # Regression problem: no activation in the last layer.
                'loss_function': 'mse',
                'optimiser': 'adam',
                'metrics': ['mae']
                }
           }

def get_model(config):
    
    params = config['model_config']
    model_name, layers, activations = params['model_name'], params['layers'], params['activations']
    loss, opt, metrics = params['loss_function'], params['optimiser'], params['metrics'] 
    
    model = Sequential(name = model_name)
    for l, (name, n_units) in enumerate(layers.items()):
        if l==0:
            model.add(Dense(units=n_units, input_dim=num_features, activation = activations[l], name = name))
        else:
            model.add(Dense(units=n_units, activation = activations[l], name = name))

    model.compile(loss=loss, optimizer=opt, metrics=metrics)
    model.summary()
    
    return model

model = get_model(config)