tensorflow2.0 公共訓練數據加載的兩個方法

學習筆記：訓練數據加載

方法一：通過tensorflow.keras.datasets加載

from tensorflow import keras

#加載fashion_mnist（時裝圖片）數據
(x_train, y_train), (x_test, y_test) = keras.datasets.fashion_mnist.load_data()

#加載mnist（手寫數字）數據
(x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data()

方法一：通過tensorflow_datasets模塊加載

首先python需要安裝tensorflow_datasets模塊

pip install tensorflow_datasets

然後程序代碼:

import tensorflow_datasets as tfds

#加載fashion_mnist（時裝圖片）數據
dataset, metadata =tfds.load('fashion_mnist', as_supervised = True, with_info = True)

#加載mnist（手寫數字）數據
dataset, metadata =tfds.load('mnist', as_supervised = True, with_info = True)

dataset_train, dataset_test = dataset['train'], dataset['test']