small_xception.py

"""Downsized version of Xception, without residual connections.
"""
from __future__ import print_function
from __future__ import absolute_import

from keras.models import Model
from keras.layers import Dense
from keras.layers import Input
from keras.layers import BatchNormalization
from keras.layers import Activation
from keras.layers import Conv2D
from keras.layers import SeparableConv2D
from keras.layers import MaxPooling2D
from keras.layers import GlobalAveragePooling2D


def SmallXception(input_shape=(200, 200, 3),
                  classes=100):
    img_input = Input(shape=input_shape)

    x = Conv2D(32, (3, 3), strides=(2, 2), use_bias=False)(img_input)
    x = BatchNormalization()(x)
    x = Activation('relu')(x)
    x = MaxPooling2D((3, 3), strides=(2, 2))(x)

    x = SeparableConv2D(128, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)
    x = Activation('relu')(x)
    x = SeparableConv2D(128, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)

    x = MaxPooling2D((3, 3), strides=(2, 2))(x)

    x = SeparableConv2D(256, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)
    x = Activation('relu')(x)
    x = SeparableConv2D(256, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)

    x = MaxPooling2D((3, 3), strides=(2, 2))(x)

    x = SeparableConv2D(512, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)
    x = Activation('relu')(x)
    x = SeparableConv2D(512, (3, 3), use_bias=False)(x)
    x = BatchNormalization()(x)

    x = GlobalAveragePooling2D(name='avg_pool')(x)
    x = Dense(classes, activation='softmax', name='predictions')(x)

    # Create model.
    model = Model(img_input, x)
    return model


if __name__ == '__main__':
    model = SmallXception()
    model.summary()