-
Notifications
You must be signed in to change notification settings - Fork 5
/
main.py
104 lines (97 loc) · 7.06 KB
/
main.py
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
# Import packages
from Modules.delete_glasses import delete_glasses
from Modules.face_extraction import smiles_extraction
from Modules.pre_processing import data_preprocessing, hog_pca_preprocessing
from Modules.test_pre_processing import test_data_preparation, test_hog_pca_preprocessing
from A1.a1 import A1
from A2.a2 import A2
from B1.b1 import B1
from B2.b2 import B2
import tensorflow as tf
# set_memory_growth() allocates exclusively the GPU memory needed
physical_devices = tf.config.experimental.list_physical_devices('GPU')
print("Num GPUs Available: ", len(physical_devices))
if len(physical_devices) is not 0:
tf.config.experimental.set_memory_growth(physical_devices[0], True)
# A1 ===================================================================================================================
# Extract smiles for A2 task before dividing all the images in 'celeba' in training and test images.
data_directory, faces_not_detected = smiles_extraction(dataset_name='celeba')
test_directory, faces_not_detected1 = smiles_extraction(dataset_name='celeba_test')
training_batches, valid_batches, test_batches = data_preprocessing(data_directory='celeba', img_size=(96, 96),
filename_column='img_name', target_column='gender',
training_percentage_size=0.85, batches_size=16,
validation_split=0.15)
input_shape = training_batches.image_shape
# Build model object.
model_A1 = A1(input_shape)
# Train model based on the training set
acc_A1_train, acc_A1_valid = model_A1.train(training_batches, valid_batches, epochs=25, verbose=2, plot=True)
# Test model based on the test set.
acc_A1_test = model_A1.test(test_batches, verbose=1, confusion_mesh=True)
# Test the model on the second larger test dataset provided
test_batches = test_data_preparation('celeba_test', filename_column='img_name', target_column='gender',
batches_size=16, img_size=(96, 96))
acc_A1_test2 = model_A1.test(test_batches, verbose=1, confusion_mesh=False)
# Clean up memory
del model_A1, physical_devices, faces_not_detected, faces_not_detected1
# A2 SVM ===============================================================================================================
x_test, x_train, x_valid, y_test, y_train, y_valid, pca, sc = hog_pca_preprocessing(dataset_name=data_directory,
img_size=(96, 48),
validation_split=0.15,
variance=0.90, training_size=0.85,
target_column='smiling')
# Build model object.
# (gamma,c) and tol values are found through grid_search.py and training_A2_plot.py (in the _Additional_code folder).
model_A2 = A2(kernel='rbf', gamma=0.001, c=0.5, tol=0.1, verbose=False)
# Train model based on the training set
acc_A2_train, acc_A2_valid = model_A2.train(x_train, x_valid, y_train, y_valid)
# Test model based on the test set.
acc_A2_test = model_A2.test(x_test, y_test, confusion_mesh=True)
# Test the model on the second larger test dataset provided
x_test, y_test = test_hog_pca_preprocessing(test_directory, pca, sc, img_size=(96, 48), target_column='smiling')
acc_A2_test2 = model_A2.test(x_test, y_test, confusion_mesh=False)
# Clean up memory
del x_test, x_train, x_valid, y_test, y_train, y_valid, data_directory, model_A2, pca, sc
# B1 ===================================================================================================================
training_batches, valid_batches, test_batches = data_preprocessing(data_directory='cartoon_set',
filename_column='file_name',
target_column='face_shape', img_size=(224, 224),
training_percentage_size=0.8, horizontal_flip=False,
batches_size=16, validation_split=0.2)
input_shape = training_batches.image_shape
# Build model object.
model_B1 = B1(input_shape)
# Train model based on the training set
acc_B1_train, acc_B1_valid = model_B1.train(training_batches, valid_batches, epochs=10, verbose=2, plot=True)
# Test model based on the test set.
acc_B1_test = model_B1.test(test_batches, verbose=1, confusion_mesh=True)
# Test the model on the second larger test dataset provided
test_batches = test_data_preparation('cartoon_set_test', filename_column='file_name', target_column='face_shape',
batches_size=16, img_size=(224, 224))
acc_B1_test2 = model_B1.test(test_batches, verbose=1, confusion_mesh=False)
# B2 ===================================================================================================================
# Execute after the B1 Task
delete_glasses(dataset_name='cartoon_set', img_size=(224, 224))
training_batches, valid_batches, test_batches = data_preprocessing(data_directory='cartoon_set',
filename_column='file_name',
target_column='eye_color', img_size=(224, 224),
training_percentage_size=0.8, batches_size=16,
horizontal_flip=False, validation_split=0.2)
input_shape = training_batches.image_shape
# Build model object.
model_B2 = B2(input_shape)
# Train model based on the training set
acc_B2_train, acc_B2_valid = model_B2.train(training_batches, valid_batches, epochs=10, verbose=2, plot=True)
# Test model based on the test set.
acc_B2_test = model_B2.test(test_batches, verbose=1, confusion_mesh=True)
# Test the model on the second larger test dataset provided
test_batches = test_data_preparation('cartoon_set_test', filename_column='file_name', target_column='eye_color',
batches_size=16, img_size=(224, 224))
acc_B2_test2 = model_B2.test(test_batches, verbose=1, confusion_mesh=False)
# RESULTS ==============================================================================================================
# Print out your results with following format:
print('Task {:<12} {:<12} {:<12} {:<12}\n'.format('Train Acc', 'Valid Acc', 'Test Acc', 'Test 2 Acc'),
'A1: {:<12.4f} {:<12.4f} {:<12.4f} {:<12.4f}\n'.format(acc_A1_train, acc_A1_valid, acc_A1_test, acc_A1_test2),
'A2: {:<12.4f} {:<12.4f} {:<12.4f} {:<12.4f}\n'.format(acc_A2_train, acc_A2_valid, acc_A2_test, acc_A2_test2),
'B1: {:<12.4f} {:<12.4f} {:<12.4f} {:<12.4f}\n'.format(acc_B1_train, acc_B1_valid, acc_B1_test, acc_B1_test2),
'B2: {:<12.4f} {:<12.4f} {:<12.4f} {:<12.4f}\n'.format(acc_B2_train, acc_B2_valid, acc_B2_test, acc_B2_test2))