keras_auto_encoder.py 文件源码

def deep_decoder1(input_shape):
    encoded = Input(shape=input_shape)
    print 'decoder input shape:', encoded._keras_shape

    batch_size = tf.shape(encoded)[0]
    x = BatchNormalization(mode=2, axis=3)(encoded)

    h, w, _ = encoded._keras_shape[1:]
    x = Deconv2D(32, 1, 1, output_shape=[batch_size, h, w, 32],
                 activation='relu', border_mode='same')(x)
    x = BatchNormalization(mode=2, axis=3)(x)
    x = Deconv2D(32, 3, 3, output_shape=[batch_size, h, w, 32],
                 activation='relu', border_mode='same')(x)
    x = BatchNormalization(mode=2, axis=3)(x)

    h *= 2; w *= 2
    x = Deconv2D(64, 3, 3, output_shape=(batch_size, h, w, 64),
                 activation='relu', border_mode='same', subsample=(2, 2))(x)
    x = BatchNormalization(mode=2, axis=3)(x)

    x = Deconv2D(64, 3, 3, output_shape=(batch_size, h, w, 64),
                 activation='relu', border_mode='same', subsample=(1, 1))(x)
    x = BatchNormalization(mode=2, axis=3)(x)

    h *= 2; w *= 2
    x = Deconv2D(32, 3, 3, output_shape=(batch_size, h, w, 32),
                 activation='relu', border_mode='same', subsample=(2, 2))(x)
    x = BatchNormalization(mode=2, axis=3)(x)
    x = Deconv2D(32, 3, 3, output_shape=(batch_size, h, w, 32),
                 activation='relu', border_mode='same', subsample=(1, 1))(x)
    x = BatchNormalization(mode=2, axis=3)(x)

    x = Deconv2D(3, 3, 3, output_shape=(batch_size, 32, 32, 3),
                 activation='linear', border_mode='same', subsample=(1, 1))(x)
    x = BatchNormalization(mode=2, axis=3)(x)
    decoded = x
    return Model(encoded, decoded)