CNN(MNIST 데이터 실습)

1. 데이터 정리

from tensorflow.keras import datasets

mnist = datasets.mnist

(X_train, y_train), (X_test, y_test) = mnist.load_data()
X_train, X_test = X_train / 255.0, X_test / 255.0

X_train = X_train.reshape((60000, 28, 28, 1))
X_test = X_test.reshape((10000, 28, 28, 1))

2. 모델

from tensorflow.keras import layers, models

model = models.Sequential([
	# 32개 특성, input_shape 차원 맞춰줌 28 * 28 * 32개의  layers -> (28, 28, 1)
    layers.Conv2D(32, kernel_size=(5, 5), strides=(1, 1), padding='same', activation='relu', input_shape=(28, 28, 1)),
    layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),
    # 줄였으니 특성을 더 잡기 64개 특성
    layers.Conv2D(64, (2,2), padding='same', activation='relu'),
    layers.MaxPooling2D(pool_size=(2,2)),
    layers.Dropout(0.25),
    # 특성 모두 펼치기
    layers.Flatten(),
    layers.Dense(1000, activation='relu'),
    layers.Dense(10, activation='softmax')
])
model.summary()
'''
Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 conv2d (Conv2D)             (None, 28, 28, 32)        832       
                                                                 
 max_pooling2d (MaxPooling2D  (None, 14, 14, 32)       0         
 )                                                               
                                                                 
 conv2d_1 (Conv2D)           (None, 14, 14, 64)        8256      
                                                                 
 max_pooling2d_1 (MaxPooling  (None, 7, 7, 64)         0         
 2D)                                                             
                                                                 
 dropout (Dropout)           (None, 7, 7, 64)          0         
                                                                 
 flatten (Flatten)           (None, 3136)              0         
                                                                 
 dense (Dense)               (None, 1000)              3137000   
                                                                 
 dense_1 (Dense)             (None, 10)                10010     
                                                                 
=================================================================
Total params: 3,156,098
Trainable params: 3,156,098
Non-trainable params: 0
'''

3. 훈련

import time

model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)
start_time = time.time()
hist = model.fit(X_train, y_train, epochs=5, verbose=1, validation_data=(X_test, y_test))
print('fit time: ', time.time() - start_time)

4. 학습 결과

• 결과가 좋다

import matplotlib.pyplot as plt
%matplotlib inline

plot_target = ['loss', 'val_loss', 'accuracy', 'val_accuracy']
plt.figure(figsize=(12, 8))
for each in plot_target:
  plt.plot(hist.history[each], label=each)

plt.legend()
plt.grid()
plt.show()

score = model.evaluate(X_test, y_test)
print('Test loss: ', score[0])
print('Test accuracy: ', score[1])
'''
313/313 [==============================] - 3s 11ms/step - loss: 0.0304 - accuracy: 0.9908
Test loss:  0.030372295528650284
Test accuracy:  0.9908000230789185
'''

5. 틀린 데이터

import numpy as np
import random

predicted_result = model.predict(X_test)
predicted_labels = np.argmax(predicted_result, axis=1)


worng_results = []

for n in range(0, len(y_test)):
  if predicted_labels[n] != y_test[n]:
    worng_results.append(n)


samples = random.sample(worng_results, 16)

plt.figure(figsize=(14, 12))

for idx, n in enumerate(samples):
  plt.subplot(4, 4, idx+1)
  plt.imshow(X_test[n].reshape(28, 28), cmap='Greys', interpolation='nearest')
  plt.title('Label: ' + str(y_test[n]) + 'Predict' + str(predicted_labels[n]))
  plt.axis('off')

plt.show()

6. 모델 저장

model.save('/content/drive/MNIST_CNN_models.hs')

Reference
1) 제로베이스 데이터스쿨 강의자료