dcgan128_model.py

# Copyright 2017 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Simple generator and discriminator models.

Based on the convolutional and "deconvolutional" models presented in
"Unsupervised Representation Learning with Deep Convolutional Generative
Adversarial Networks" by A. Radford et. al.

Extended to:
- Support 128 * 128 images
- Add 2 additional layers to discriminator and generator.
- Add support for image_size flags. Note: You will need to modify D and G to support new image sizes.
"""

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import tensorflow as tf

image_size = 128
s16 = image_size // 32
df_dim = 64  # Number of filters for discriminator
gf_dim = 64  # Number of filters for generator
c_dim = 3  # Number of image channels


def _leaky_relu(x):
    return tf.nn.leaky_relu(x, alpha=0.2)


def _batch_norm(x, is_training, name):
    return tf.layers.batch_normalization(
        x, momentum=0.9, epsilon=1e-5, training=is_training, name=name)


def _dense(x, channels, name):
    return tf.layers.dense(
        x, channels,
        kernel_initializer=tf.truncated_normal_initializer(stddev=0.02),
        name=name)


def _conv2d(x, filters, kernel_size, stride, name):
    return tf.layers.conv2d(
        x, filters, [kernel_size, kernel_size],
        strides=[stride, stride], padding='same',
        kernel_initializer=tf.truncated_normal_initializer(stddev=0.02),
        name=name)


def _deconv2d(x, filters, kernel_size, stride, name):
    return tf.layers.conv2d_transpose(
        x, filters, [kernel_size, kernel_size],
        strides=[stride, stride], padding='same',
        kernel_initializer=tf.truncated_normal_initializer(stddev=0.02),
        name=name)


def discriminator(x, is_training=True, scope='Discriminator'):
    with tf.variable_scope(scope, reuse=tf.AUTO_REUSE):
        x = _conv2d(x, df_dim, 5, 2, name='d_conv1')
        x = _leaky_relu(x)

        x = _conv2d(x, df_dim*2, 5, 2, name='d_conv2')
        x = _leaky_relu(_batch_norm(x, is_training, name='d_bn2'))

        x = _conv2d(x, df_dim*4, 5, 2, name='d_conv3')
        x = _leaky_relu(_batch_norm(x, is_training, name='d_bn3'))

        x = _conv2d(x, df_dim*8, 5, 2, name='d_conv4')
        x = _leaky_relu(_batch_norm(x, is_training, name='d_bn4'))

        x = _conv2d(x, df_dim*16, 5, 2, name='d_conv5')
        x = _leaky_relu(_batch_norm(x, is_training, name='d_bn5'))

        x = tf.reshape(x, [-1, s16, s16, df_dim*16])

        x = _dense(x, 1, name='d_fc_4')

        return x


def generator(x, is_training=True, scope='Generator'):
    with tf.variable_scope(scope, reuse=tf.AUTO_REUSE):
        x = _dense(x, gf_dim * 16 * s16 * s16, name='g_fc1')
        x = tf.nn.relu(_batch_norm(x, is_training, name='g_bn1'))

        x = tf.reshape(x, [-1, s16, s16, gf_dim*16], name="reshape_1")

        x = _deconv2d(x, gf_dim * 8, 5, 2, name='g_dconv1')
        x = tf.nn.relu(_batch_norm(x, is_training, name='g_bn1_5'))

        x = _deconv2d(x, gf_dim * 4, 5, 2, name='g_dconv2')
        x = tf.nn.relu(_batch_norm(x, is_training, name='g_bn2'))

        x = _deconv2d(x, gf_dim * 2, 5, 2, name='g_dconv3')
        x = tf.nn.relu(_batch_norm(x, is_training, name='g_bn3'))

        x = _deconv2d(x, gf_dim, 5, 2, name='g_dconv4')
        x = tf.nn.relu(_batch_norm(x, is_training, name='g_bn4'))

        x = _deconv2d(x, c_dim, 5, 2, name='g_dconv5')
        x = tf.tanh(x)

        return x