Gradient-Ascent_FK-GRiD/encoder_training_script.py at master · archana1998/Gradient-Ascent_FK-GRiD · GitHub

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
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
import numpy as np
import matplotlib.pyplot as plt

from keras.layers import Input, Dense, Conv2D, MaxPooling2D, UpSampling2D, Flatten
from keras.models import Model
from keras import backend as K
import tensorflow.keras.layers
import keras

#Defining the input image size - chosen so it can be recreated easily
input_img = Input(shape=(320, 192, 3))

#Model architecture definition
#Consists of convultional layers to help learn the features of the clothing item - sleeve type, striped, collar type, etc
x = Conv2D(64, (3, 3), activation='relu', padding='same')(input_img)
x = MaxPooling2D((2, 2), padding='same')(x)
x = Conv2D(32, (3, 3), activation='relu', padding='same')(x)
x = MaxPooling2D((2, 2), padding='same')(x)
x = Conv2D(16, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = MaxPooling2D((2, 2), padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
encoded = MaxPooling2D((2, 2), padding='same')(x)

x = Conv2D(8, (3, 3), activation='relu', padding='same')(encoded)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = UpSampling2D((2, 2))(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)
x = Conv2D(16, (3, 3), activation='relu', padding='same')(x)
x = UpSampling2D((2, 2))(x)
x = Conv2D(32, (3, 3), activation='relu', padding='same')(x)
x = UpSampling2D((2, 2))(x)
x = Conv2D(64, (3, 3), activation='relu', padding='same')(x)
x = UpSampling2D((2, 2))(x)
decoded = Conv2D(3, (3, 3), activation='sigmoid', padding='same')(x)

autoencoder = Model(input_img, decoded)

optimizer = keras.optimizers.Adam(lr=0.001)
autoencoder.compile(optimizer=optimizer, loss='binary_crossentropy')

import zipfile

zip_ref = zipfile.ZipFile("Zipped_final.zip", 'r')
zip_ref.extractall("/tmp")
zip_ref.close()


#Loading in the data and resizing the images
from skimage.io import imread_collection

col_dir = '/tmp/Zipped/Amazon Images/*.jpg'
col_amazon = imread_collection(col_dir)
col_amazon = list(col_amazon)

from skimage.transform import resize

for i in range(len(col_amazon)):
    col_amazon[i] = resize(col_amazon[i], (320, 192, 3))

col_dir = '/tmp/Zipped/Flipkart Images/*.jpg'
col_fk = imread_collection(col_dir)
col_fk = list(col_fk)

for i in range(len(col_fk)):
    col_fk[i] = resize(col_fk[i], (320, 192, 3))

col_dir = '/tmp/Zipped/Myntra Images/*.jpg'
col_myntra = imread_collection(col_dir)
col_myntra = list(col_myntra)

for i in range(len(col_myntra)):
    col_myntra[i] = resize(col_myntra[i], (320, 192, 3))

col_dir = '/tmp/Zipped/Pinterest Men Images/*.jpg'
col_pinmen = imread_collection(col_dir)
col_pinmen = list(col_pinmen)

for i in range(len(col_pinmen)):
    col_pinmen[i] = resize(col_pinmen[i], (320, 192, 3))

col_dir = '/tmp/Zipped/Pinterest Women Images/*.jpg'
col_pinwom = imread_collection(col_dir)
col_pinwom = list(col_pinwom)

for i in range(len(col_pinwom)):
    col_pinwom[i] = resize(col_pinwom[i], (320, 192, 3))

col_dir = '/tmp/Zipped/Vogue Images/*.jpg'
col_vog = imread_collection(col_dir)
col_vog = list(col_vog)

for i in range(len(col_vog)):
    col_vog[i] = resize(col_vog[i], (320, 192, 3))

col_clothes = col_amazon+col_fk+col_pinmen+col_vog+col_myntra+col_pinwom

x_train = np.array(col_clothes[0:3000])
x_test = np.array(col_clothes[3000:])
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')

#Training the model
autoencoder.fit(x_train, x_train,
                epochs=50,
                batch_size=32,
                validation_data=(x_test, x_test))

#Saving the model
encoder = Model(input_img, encoded)
encoder.save("encoder")