CNN MNIST손글씨 데이터 학습

In [1]:

from IPython.core.display import display, HTML
display(HTML("<style>.container {width:90% !important;}</style>"))

라이브러리 불러오기¶

In [2]:

import tensorflow as tf
from tensorflow.keras import datasets, layers, models

import numpy as np
import matplotlib.pyplot as plt
import cv2 as cv
import os

print(tf.__version__)

2.3.1

데이터셋 불러오기 및 정규화¶

In [3]:

(train_images, train_labels), (test_images, test_labels) = datasets.mnist.load_data()
train_images = train_images.reshape((60000, 28, 28, 1))
test_images = test_images.reshape((10000, 28, 28, 1))

# 픽셀 값을 0~1 사이로 정규화합니다.
train_images, test_images = train_images / 255.0, test_images / 255.0

CNN 모델 생성¶

In [4]:

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))

model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(10, activation='softmax'))
model.summary()

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 26, 26, 32)        320       
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 13, 13, 32)        0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 11, 11, 64)        18496     
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 5, 5, 64)          0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 3, 3, 64)          36928     
_________________________________________________________________
flatten (Flatten)            (None, 576)               0         
_________________________________________________________________
dense (Dense)                (None, 64)                36928     
_________________________________________________________________
dense_1 (Dense)              (None, 10)                650       
=================================================================
Total params: 93,322
Trainable params: 93,322
Non-trainable params: 0
_________________________________________________________________

모델 컴파일 및 훈련¶

In [5]:

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

model.fit(train_images, train_labels, epochs=5)

Epoch 1/5
1875/1875 [==============================] - 17s 9ms/step - loss: 0.1464 - accuracy: 0.9551
Epoch 2/5
1875/1875 [==============================] - 17s 9ms/step - loss: 0.0460 - accuracy: 0.9853
Epoch 3/5
1875/1875 [==============================] - 17s 9ms/step - loss: 0.0331 - accuracy: 0.9896
Epoch 4/5
1875/1875 [==============================] - 17s 9ms/step - loss: 0.0261 - accuracy: 0.9920
Epoch 5/5
1875/1875 [==============================] - 17s 9ms/step - loss: 0.0205 - accuracy: 0.9933

Out[5]:

<tensorflow.python.keras.callbacks.History at 0x1d090408e50>

모델 평가¶

In [6]:

test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)

313/313 - 1s - loss: 0.0313 - accuracy: 0.9908

In [7]:

print(test_acc)

0.9908000230789185

모델 저장¶

In [8]:

model.save('mnist_cnn_model.h5')