food_recognition_resnet.py

from custom_layers import Segmentation
from utils import *
import tensorflow as tf
from keras.models import Sequential
from tensorflow.keras.applications import ResNet50


# Initialize the Pretrained Model
def get_model():
    feature_extractor = ResNet50(weights='imagenet',
                                 input_shape=(WIDTH, HEIGHT, 3),
                                 include_top=False)

    num_layers = len(feature_extractor.layers)
    for layer in feature_extractor.layers[:num_layers // 2]:
        layer.trainable = False

    model = Sequential()
    model.add(tf.keras.Input(shape=(WIDTH, HEIGHT, 3)))
    # model.add(Segmentation())
    model.add(feature_extractor)
    model.add(tf.keras.layers.GlobalAveragePooling2D())
    model.add(tf.keras.layers.Dropout(0.5))
    model.add(tf.keras.layers.BatchNormalization())
    model.add(tf.keras.layers.Dense(NUM_CLASSES, activation='softmax'))

    # Compile it
    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

    # Print The Summary of The Model
    model.summary()
    return model


if __name__ == '__main__':
    model = get_model()

    train_xy, validation_xy, test_x = get_data_generators()
    checkpoint_filepath = 'training\\weights.{epoch:02d}-{val_loss:.2f}.hdf5'
    Path(checkpoint_filepath.split('\\')[0]).mkdir(parents=True, exist_ok=True)

    model_checkpoint_callback = tf.keras.callbacks.ModelCheckpoint(
        filepath=checkpoint_filepath,
        save_weights_only=True,
        monitor='val_accuracy',
        mode='max',
        save_best_only=True)

    model.fit(train_xy, epochs=30, validation_data=validation_xy, callbacks=[model_checkpoint_callback])