[딥러닝] 파이토치(pytorch)로 MNIST 분류(classification) 간단히 구현하기

1. 필요한 라이브러리 당기기 (pytorch, mnist 라이브러리)

import torch
from tensorflow.keras.datasets import mnist

2. mnist 다운로드 및 train, test 데이터 저장

(train_images, train_labels), (test_images, test_labels) = mnist.load_data()
print(train_images.shape, train_labels.shape)
# ((60000, 28, 28), (60000,) - 28*28 픽셀 이미지 60000장(train data)
print(test_images.shape, test_labels.shape)
# (10000, 28, 28), (10000,)) - 28*28 픽셀 이미지 10000장(test data)
print(type(train_images), type(train_labels), type(test_images), type(test_labels))
# (numpy.ndarray, numpy.ndarray, numpy.ndarray, numpy.ndarray) - numpy ndarray 타입

3. 2차원인 이미지를 1차원으로 변경

train_images = train_images.reshape((60000, 28*28))
# (60000, 28, 28) -> (60000, 784) - 1차원으로 변경
test_images = test_images.reshape((10000, 28*28))
# (10000, 28, 28) -> (10000, 784)