Understanding Deepfakes with Keras

Posted on | In , | |

Understanding Deepfakes with Keras

DCGAN

Task 1: Importing Libraries and Helper Functions

Please note: If you haven’t already, please install the required packages by executing the code cell below.

1
2
# !pip3 install tensorflow==2.1.0 pillow matplotlib
# !pip3 install git+https://github.com/am1tyadav/tfutils.git
1
2
3
4
5
6
7
8
9
10
11
12
13
14
%matplotlib notebook

import tensorflow as tf
import numpy as np
import os
import tfutils

from matplotlib import pyplot as plt
from tensorflow.keras.layers import Dense, Flatten, Conv2D, BatchNormalization
from tensorflow.keras.layers import Conv2DTranspose, Reshape, LeakyReLU
from tensorflow.keras.models import Model, Sequential
from PIL import Image

print('TensorFlow version:', tf.__version__)
TensorFlow version: 2.1.0

Task 2: Importing and Plotting the Data

1
2
3
4
5
6
(x_train, y_train), (x_test, y_test) = tfutils.datasets.mnist.load_data(one_hot=False)

x_train = tfutils.datasets.mnist.load_subset([0], x_train, y_train)
x_test = tfutils.datasets.mnist.load_subset([0], x_test, y_test)

x = np.concatenate([x_train, x_test], axis=0)
1
x.shape
(6903, 784)

1
tfutils.datasets.mnist.plot_ten_random_examples(plt, x, np.zeros((x.shape[0], 1))).show()
<IPython.core.display.Javascript object>

Task 3: Discriminator

Artist and Critic

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
size = 28
noise_dim = 1

discriminator = Sequential([
    Conv2D(64, 3, strides=2, input_shape=(28, 28, 1)),
    LeakyReLU(),
    BatchNormalization(),
    
    Conv2D(128, 5, strides=2),
    LeakyReLU(),
    BatchNormalization(),
    
    Conv2D(256, 5, strides=2),
    LeakyReLU(),
    BatchNormalization(),
    
    Flatten(),
    Dense(1, activation='sigmoid')
])

opt = tf.keras.optimizers.Adam(lr=2e-4, beta_1=0.5)

discriminator.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])
discriminator.summary()
Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 13, 13, 64)        640       
_________________________________________________________________
leaky_re_lu (LeakyReLU)      (None, 13, 13, 64)        0         
_________________________________________________________________
batch_normalization (BatchNo (None, 13, 13, 64)        256       
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 5, 5, 128)         204928    
_________________________________________________________________
leaky_re_lu_1 (LeakyReLU)    (None, 5, 5, 128)         0         
_________________________________________________________________
batch_normalization_1 (Batch (None, 5, 5, 128)         512       
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 1, 1, 256)         819456    
_________________________________________________________________
leaky_re_lu_2 (LeakyReLU)    (None, 1, 1, 256)         0         
_________________________________________________________________
batch_normalization_2 (Batch (None, 1, 1, 256)         1024      
_________________________________________________________________
flatten (Flatten)            (None, 256)               0         
_________________________________________________________________
dense (Dense)                (None, 1)                 257       
=================================================================
Total params: 1,027,073
Trainable params: 1,026,177
Non-trainable params: 896
_________________________________________________________________

Task 4: Generator

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
generator = Sequential([
    Dense(256, activation='relu', input_shape=(noise_dim,)),
    Reshape((1, 1, 256)),
    
    Conv2DTranspose(256, 5, activation='relu'),
    BatchNormalization(),
    Conv2DTranspose(128, 5, activation='relu'),
    BatchNormalization(),

    Conv2DTranspose(64, 5, strides=2, activation='relu'),
    BatchNormalization(),
    Conv2DTranspose(32, 5, activation='relu'),
    BatchNormalization(),

    Conv2DTranspose(1, 4, activation='sigmoid')

])

generator.summary()
Model: "sequential_1"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_1 (Dense)              (None, 256)               512       
_________________________________________________________________
reshape (Reshape)            (None, 1, 1, 256)         0         
_________________________________________________________________
conv2d_transpose (Conv2DTran (None, 5, 5, 256)         1638656   
_________________________________________________________________
batch_normalization_3 (Batch (None, 5, 5, 256)         1024      
_________________________________________________________________
conv2d_transpose_1 (Conv2DTr (None, 9, 9, 128)         819328    
_________________________________________________________________
batch_normalization_4 (Batch (None, 9, 9, 128)         512       
_________________________________________________________________
conv2d_transpose_2 (Conv2DTr (None, 21, 21, 64)        204864    
_________________________________________________________________
batch_normalization_5 (Batch (None, 21, 21, 64)        256       
_________________________________________________________________
conv2d_transpose_3 (Conv2DTr (None, 25, 25, 32)        51232     
_________________________________________________________________
batch_normalization_6 (Batch (None, 25, 25, 32)        128       
_________________________________________________________________
conv2d_transpose_4 (Conv2DTr (None, 28, 28, 1)         513       
=================================================================
Total params: 2,717,025
Trainable params: 2,716,065
Non-trainable params: 960
_________________________________________________________________

1
2
3
4
5
noise = np.random.randn(1, noise_dim)
gen_image = generator.predict(noise)[0]

plt.figure()
plt.imshow(np.reshape(gen_image, (28, 28)), cmap='binary');
<IPython.core.display.Javascript object>

Task 5: Generative Adversarial Network (GAN)

1
2
3
4
5
6
7
8
9
10
11
12
input_layer = tf.keras.layers.Input(shape=(noise_dim,))
gen_out = generator(input_layer)
disc_out = discriminator(gen_out)

gan = Model(
    input_layer,
    disc_out
)

discriminator.trainable = False
gan.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])
gan.summary()
Model: "model"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_1 (InputLayer)         [(None, 1)]               0         
_________________________________________________________________
sequential_1 (Sequential)    (None, 28, 28, 1)         2717025   
_________________________________________________________________
sequential (Sequential)      (None, 1)                 1027073   
=================================================================
Total params: 3,744,098
Trainable params: 2,716,065
Non-trainable params: 1,028,033
_________________________________________________________________

Tasks 6 and 7: Training the GAN

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
%%time

epochs = 25
batch_size = 128
steps_per_epoch = int(2 * x.shape[0]/batch_size)

print('Steps per epoch=', steps_per_epoch)

dp = tfutils.plotting.DynamicPlot(plt, 5, 5, (8, 8))

for e in range(0, epochs):
    
    dp.start_of_epoch(e)
    
    for step in range(0, steps_per_epoch):
        true_examples = x[int(batch_size/2)*step: int(batch_size/2)*(step + 1)]
        true_examples = np.reshape(true_examples, (true_examples.shape[0], 28, 28, 1))

        noise = np.random.randn(int(batch_size/2), noise_dim)
        generated_examples = generator.predict(noise)

        x_batch = np.concatenate([generated_examples, true_examples], axis=0)
        y_batch = np.array([0] * int(batch_size/2) + [1] * int(batch_size/2))

        indices = np.random.choice(range(batch_size), batch_size, replace=False)
        x_batch = x_batch[indices]
        y_batch = y_batch[indices]

        # train the discriminator
        discriminator.trainable = True
        discriminator.train_on_batch(x_batch, y_batch)
        discriminator.trainable = False

        # train the generator
        loss, _ = gan.train_on_batch(noise, np.ones((int(batch_size/2), 1)))

        _, acc = discriminator.evaluate(x_batch, y_batch, verbose=False)

    noise = np.random.randn(1, noise_dim)
    generated_example = generator.predict(noise)[0]
    
    dp.end_of_epoch(np.reshape(generated_example, (28, 28)), 'binary',
                   'DiscAcc:{:.2f}'.format(acc), 'GANLoss:{:.2f}'.format(loss))
Steps per epoch= 107



<IPython.core.display.Javascript object>


1
2