DCGAN how to go RGB instead of greyscale-CodePudding

I have this DCGAN that is pretty close to the TensorFlow docs.

Here is the tutorial: https://www.tensorflow.org/tutorials/generative/dcgan

It uses greyscale values in the test data. I am looking to start training with color data instead of just black and white.

I am assuming that the shape of the training data will need to change, but does the shape of the generator model need to change too?

How can I adapt this code to an RGB implementation?

from google.colab import drive

drive.mount('/content/drive')

import tensorflow as tf

import glob
import matplotlib.pyplot as plt
import numpy as np
import os
import PIL
from tensorflow.keras import layers
import time

from IPython import display

train_dataset = tf.keras.preprocessing.image_dataset_from_directory(
  "/content/drive/MyDrive/birds",
  seed=123,
  validation_split=0,
  image_size=(112, 112),
  color_mode="grayscale",
  shuffle=True,
  batch_size=1)

train_images_array = []
for images, _ in train_dataset:
    for i in range(len(images)):
      train_images_array.append(images[i])
      

train_images = np.array(train_images_array)
train_images = train_images.reshape(train_images.shape[0],112,112,1).astype('float32')

train_images = (train_images - 127.5) / 127.5  # Normalize the images to [-1, 1]

BUFFER_SIZE = 60000
BATCH_SIZE = 8

# Batch and shuffle the data
dataset_ = tf.data.Dataset.from_tensor_slices(train_images).shuffle(BUFFER_SIZE).batch(BATCH_SIZE)

def make_generator_model():
    model = tf.keras.Sequential()
    model.add(layers.Dense(7*7*256, use_bias=False, input_shape=(100,)))
    model.add(layers.BatchNormalization())
    model.add(layers.LeakyReLU())

    model.add(layers.Reshape((7, 7, 256)))
    assert model.output_shape == (None, 7, 7, 256)  # Note: None is the batch size

    model.add(layers.Conv2DTranspose(128, (5, 5), strides=(1, 1), padding='same', use_bias=False))
    assert model.output_shape == (None, 7, 7, 128)
    model.add(layers.BatchNormalization())
    model.add(layers.LeakyReLU())

    model.add(layers.Conv2DTranspose(64, (5, 5), strides=(2, 2), padding='same', use_bias=False))
    assert model.output_shape == (None, 14, 14, 64)
    model.add(layers.BatchNormalization())
    model.add(layers.LeakyReLU())

    model.add(layers.Conv2DTranspose(1, (20, 20), strides=(8, 8), padding='same', use_bias=False, activation='tanh'))
    assert model.output_shape == (None, 112, 112, 1)
    return model

generator = make_generator_model()

noise = tf.random.normal([1, 100])
generated_image = generator(noise, training=False)

plt.imshow(generated_image[0, :, :, 0], cmap='gray')

def make_discriminator_model():
    model = tf.keras.Sequential()
    model.add(layers.Conv2D(64, (10, 10), strides=(2, 2), padding='same', input_shape=[112, 112, 1]))
    model.add(layers.LeakyReLU())
    model.add(layers.Dropout(0.3))

    model.add(layers.Conv2D(128, (5, 5), strides=(2, 2), padding='same', input_shape=[112, 112, 1]))
    model.add(layers.LeakyReLU())
    model.add(layers.Dropout(0.3))

    model.add(layers.Flatten())
    model.add(layers.Dense(1))

    return model

discriminator = make_discriminator_model()
decision = discriminator(generated_image)
print (decision)

# This method returns a helper function to compute cross entropy loss
cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)

def discriminator_loss(real_output, fake_output):
    real_loss = cross_entropy(tf.ones_like(real_output), real_output)
    fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)
    total_loss = real_loss   fake_loss
    return total_loss

def generator_loss(fake_output):
    return cross_entropy(tf.ones_like(fake_output), fake_output)

generator_optimizer = tf.keras.optimizers.Adam(1e-4)
discriminator_optimizer = tf.keras.optimizers.Adam(1e-4)

checkpoint_dir = '/content/drive/MyDrive/training_checkpoints11'
checkpoint_prefix = os.path.join(checkpoint_dir, "ckpt")
checkpoint = tf.train.Checkpoint(generator_optimizer=generator_optimizer,
                                 discriminator_optimizer=discriminator_optimizer,
                                 generator=generator,
                                 discriminator=discriminator)

EPOCHS = 50
noise_dim = 100
num_examples_to_generate = 16

# You will reuse this seed overtime (so it's easier)
# to visualize progress in the animated GIF)
seed = tf.random.normal([num_examples_to_generate, noise_dim])



def generate_and_save_images(model, epoch, test_input):
  # Notice `training` is set to False.
  # This is so all layers run in inference mode (batchnorm).
  predictions = model(test_input, training=False)

  fig = plt.figure(figsize=(4, 4))

  for i in range(predictions.shape[0]):
      plt.subplot(4, 4, i 1)
      plt.imshow(predictions[i, :, :, 0] * 127.5   127.5, cmap='gray')
      plt.axis('off')

  plt.savefig('image_at_epoch_{:04d}.png'.format(epoch))

  plt.show()

# Notice the use of `tf.function`
# This annotation causes the function to be "compiled".
@tf.function
def train_step(images):
    noise = tf.random.normal([BATCH_SIZE, noise_dim])

    with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:
      generated_images = generator(noise, training=True)

      real_output = discriminator(images, training=True)
      fake_output = discriminator(generated_images, training=True)

      gen_loss = generator_loss(fake_output)
      disc_loss = discriminator_loss(real_output, fake_output)

    gradients_of_generator = gen_tape.gradient(gen_loss, generator.trainable_variables)
    gradients_of_discriminator = disc_tape.gradient(disc_loss, discriminator.trainable_variables)

    generator_optimizer.apply_gradients(zip(gradients_of_generator, generator.trainable_variables))
    discriminator_optimizer.apply_gradients(zip(gradients_of_discriminator, discriminator.trainable_variables))

def train(dataset, epochs):
  for epoch in range(epochs):
    start = time.time()
    for image_batch in dataset:
      train_step(image_batch)

    # Produce images for the GIF as you go
    display.clear_output(wait=True)
    generate_and_save_images(generator,
                             epoch   1,
                             seed)

    # Save the model every 1 epochs
    if (epoch   1) % 8 == 0:
      checkpoint.save(file_prefix = checkpoint_prefix)

    print ('Time for epoch {} is {} sec'.format(epoch   1, time.time()-start))

  # Generate after the final epoch
  display.clear_output(wait=True)
  generate_and_save_images(generator,
                          epochs,
                           seed)
  return

train(dataset_, 128)

noise = tf.random.normal([1, 100])
generated_image = generator(noise, training=False)
print(generated_image.shape)

plt.imshow(generated_image[0, :, :, 0], cmap='gray')


checkpoint.restore(tf.train.latest_checkpoint(checkpoint_dir))

CodePudding user response：

Yes the generator needs to be changed too. Greyscale has one channel and you need three.

So you need to change

    model.add(layers.Conv2DTranspose(1, (20, 20), strides=(8, 8), padding='same', use_bias=False, activation='tanh'))
    assert model.output_shape == (None, 112, 112, 1)

    model.add(layers.Conv2DTranspose(3, (20, 20), strides=(8, 8), padding='same', use_bias=False, activation='tanh'))
    assert model.output_shape == (None, 112, 112, 3)