aboutsummaryrefslogtreecommitdiff
path: root/lenet.py
blob: a94259bab516cd66e93f9b84cb2f499fd535584d (plain)
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
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
from __future__ import print_function
import tensorflow.keras as keras
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Flatten
from tensorflow.keras.layers import Conv2D, MaxPooling2D, AveragePooling2D
from tensorflow.keras import backend as K
from tensorflow.keras import optimizers
import matplotlib.pyplot as plt
from tensorflow.keras.metrics import categorical_accuracy
import numpy as np
import random
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from sklearn.model_selection import train_test_split
from sklearn.decomposition import PCA
from classifier_metrics_impl import classifier_score_from_logits
from sklearn.utils import shuffle
from sklearn.manifold import TSNE
import scikitplot as skplt

def import_mnist():
  from tensorflow.examples.tutorials.mnist import input_data
  mnist = input_data.read_data_sets("MNIST_data/", reshape=False)
  X_train, y_train = mnist.train.images, mnist.train.labels
  X_validation, y_validation = mnist.validation.images, mnist.validation.labels
  X_test, y_test = mnist.test.images, mnist.test.labels
  y_train = keras.utils.to_categorical(y_train, 10)
  y_validation = keras.utils.to_categorical(y_validation, 10)  
  y_test = keras.utils.to_categorical(y_test, 10)
  return X_train, y_train, X_validation, y_validation, X_test, y_test

def plot_images(images, cls_true, cls_pred=None):
    assert len(images) == len(cls_true) == 9
    img_shape = (32, 32)
    # Create figure with 3x3 sub-plots.
    fig, axes = plt.subplots(3, 3)
    fig.subplots_adjust(hspace=0.3, wspace=0.3)
    for i, ax in enumerate(axes.flat):
        # Plot image.
        ax.imshow(images[i].reshape(img_shape), cmap='binary')
        # Show true and predicted classes.
        if cls_pred is None:
            xlabel = "True: {0}".format(cls_true[i])
        else:
            xlabel = "True: {0}, Pred: {1}".format(cls_true[i], cls_pred[i])
        ax.set_xlabel(xlabel)        
        ax.set_xticks([])
        ax.set_yticks([])  
    plt.show()

def plot_example_errors(y_pred, y_true, X_test):
    correct_prediction = np.equal(y_pred, y_true)
    incorrect = np.equal(correct_prediction, False)
    images = X_test[incorrect]
    cls_pred = y_pred[incorrect]
    cls_true = y_true[incorrect]
    plot_images(images=images[0:9], cls_true=cls_true[0:9], cls_pred=cls_pred[0:9].astype(np.int))

def get_lenet(shape):
  model = keras.Sequential()
  model.add(Conv2D(filters=6, kernel_size=(3, 3), activation='relu', input_shape=shape))
  model.add(AveragePooling2D())

  model.add(Conv2D(filters=16, kernel_size=(3, 3), activation='relu'))
  model.add(AveragePooling2D())
  model.add(Flatten())
  
  model.add(Dense(units=120, activation='relu'))
  model.add(Dense(units=84, activation='relu'))
  model.add(Dense(units=10, activation = 'softmax'))
  return model

def get_lenet_icp(shape):
    model = keras.Sequential()
    model.add(Conv2D(filters=6, kernel_size=(3, 3), activation='relu', input_shape=(32,32,1)))
    model.add(AveragePooling2D())
      
    model.add(Conv2D(filters=16, kernel_size=(3, 3), activation='relu'))
    model.add(AveragePooling2D())
    model.add(Flatten())

    model.add(Dense(units=120, activation='relu'))
    model.add(Dense(units=84, activation='relu'))
    model.add(Dense(units=10, activation = 'relu'))
    return model

def plot_history(history, metric = None):
  # Plots the loss history of training and validation (if existing)
  # and a given metric
  
  if metric != None:
    fig, axes = plt.subplots(2,1)
    axes[0].plot(history.history[metric])
    try:
      axes[0].plot(history.history['val_'+metric])
      axes[0].legend(['Train', 'Val'])
    except:
      pass
    axes[0].set_title('{:s}'.format(metric))
    axes[0].set_ylabel('{:s}'.format(metric))
    axes[0].set_xlabel('Epoch')
    fig.subplots_adjust(hspace=0.5)
    axes[1].plot(history.history['loss'])
    try:
      axes[1].plot(history.history['val_loss'])
      axes[1].legend(['Train', 'Val'])
    except:
      pass
    axes[1].set_title('Model Loss')
    axes[1].set_ylabel('Loss')
    axes[1].set_xlabel('Epoch')
  else:
    plt.plot(history.history['loss'])
    try:
      plt.plot(history.history['val_loss'])
      plt.legend(['Train', 'Val'])
    except:
      pass
    plt.title('Model Loss')
    plt.ylabel('Loss')
    plt.xlabel('Epoch')

def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100, metrics=[categorical_accuracy], optimizer = None, keep_training = False, verbose=1):
  shape = (32, 32, 1)

  # Pad data to 32x32 (MNIST is 28x28)
  x_train = np.pad(x_train, ((0,0),(2,2),(2,2),(0,0)), 'constant')
  x_val = np.pad(x_val, ((0,0),(2,2),(2,2),(0,0)), 'constant')

  model = get_lenet(shape)

  if optimizer == None:
      optimizer = optimizers.SGD(lr=0.001, decay=1e-6, momentum=0.9, nesterov=True)

  model.compile(loss='categorical_crossentropy', metrics=metrics, optimizer=optimizer)
  if keep_training:
    model.load_weights('./weights.h5')
  history = model.fit(x_train, y_train, batch_size=batch_size, epochs=epochs, verbose=verbose, validation_data = (x_val, y_val))
  model.save_weights('./model_gan.h5')
  plot_history(history, 'categorical_accuracy')
  plot_history(history)
  model.save_weights('./weights.h5')
  return model 

def plot_probas(model, x_test, y_true):
    y_true = np.argmax(y_true, axis=1)
    x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
    probas = model.predict(x_test)
    skplt.metrics.plot_roc(y_true, probas)
    plt.show()
    skplt.metrics.plot_precision_recall_curve(y_true, probas)
    plt.show()

def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False):
  x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
  logits = model.predict(x_test)
  tf_logits = tf.convert_to_tensor(logits, dtype=tf.float32)
  inception_score = tf.keras.backend.eval(classifier_score_from_logits(tf_logits))
  y_pred = np.argmax(logits, axis=1)
  y_true = np.argmax(y_true, axis=1)
  plot_example_errors(y_pred, y_true, x_test)
  cm = confusion_matrix(y_true, y_pred)
  if conf_mat:
    plt.matshow(cm, cmap='Blues')
    plt.colorbar()
    plt.ylabel('Actual')
    plt.xlabel('Predicted')
    plt.show()
  if pca:
    set_pca = PCA(n_components=2)
    pca_rep = set_pca.fit_transform(logits)
    pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
    plt.scatter(pca_rep[:5000, 0], pca_rep[:5000, 1], c=y_tmp[:5000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
    plt.xlabel('Feature 1')
    plt.ylabel('Feature 2')
    plt.colorbar();
    plt.show()
  if tsne:
    tsne = TSNE(n_components=2, random_state=0)
    components = tsne.fit_transform(logits)
    print(components.shape)
    components, y_tmp = shuffle(components, y_true, random_state=0)
    plt.scatter(components[:5000, 0], components[:5000, 1], c=y_tmp[:5000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
    plt.xlabel('Feature 1')
    plt.ylabel('Feature 2')
    plt.colorbar();
    plt.show()

  
  return accuracy_score(y_true, y_pred), inception_score

def mix_data(X_train, y_train, X_validation, y_validation, train_gen, tr_labels_gen, val_gen, val_labels_gen, split=0):

  if split == 0:
    train_data = X_train
    train_labels = y_train
    val_data = X_validation
    val_labels = y_validation
  
  elif split == 1:
    train_data = train_gen
    train_labels = tr_labels_gen
    val_data = val_gen
    val_labels = val_labels_gen
  
  else:
    X_train_gen, _, y_train_gen, _ = train_test_split(train_gen, tr_labels_gen, test_size=1-split, random_state=0, stratify=tr_labels_gen)
    X_train_original, _, y_train_original, _ = train_test_split(X_train, y_train, test_size=split, random_state=0, stratify=y_train)
    X_validation_gen, _, y_validation_gen, _ = train_test_split(val_gen, val_labels_gen, test_size=1-split, random_state=0, stratify=val_labels_gen)
    X_validation_original, _, y_validation_original, _ = train_test_split(X_validation, y_validation, test_size=split, random_state=0, stratify=y_validation)
    train_data = np.concatenate((X_train_gen, X_train_original), axis=0)
    train_labels = np.concatenate((y_train_gen, y_train_original), axis=0)
    val_data = np.concatenate((X_validation_gen, X_validation_original), axis=0)
    val_labels = np.concatenate((y_validation_gen, y_validation_original), axis=0)

  return train_data, train_labels, val_data, val_labels

# If file run directly, perform quick test
if __name__ == '__main__':
  x_train, y_train, x_val, y_val, x_t, y_t = import_mnist()
  print(y_t.shape)
  model = train_classifier(x_train[:100], y_train[:100], x_val, y_val, epochs=3)
  print(test_classifier(model, x_t, y_t, pca=False, tsne=True))