Two ways of tensorflow2 training data sets

Mode one ：

def pre_process(x, y):
    x = 2. * tf.cast(x, dtype=tf.float32) / 255. - 1.
    y = tf.cast(y, dtype=tf.int32)
    return x, y


(x_train, y_train), (x_test, y_test) = tf.keras.datasets.cifar100.load_data()
x_train, y_train = pre_process(x_train, y_train)
x_test, y_test = pre_process(x_test, y_test)
print(x_train.shape, y_train.shape)

history = net.fit(x_train, y_train,
                  batch_size=512,
                  epochs=100,
                  validation_split=0.2)

test_scores = net.evaluate(x_test, y_test, verbose=2)

Training mode II ：

def pre_process(x, y):
    # [0,255] => [-1,1] ,[-1,1] It may be the most suitable range for neural network calculation 
    x = 2. * tf.cast(x, dtype=tf.float32) / 255. - 1
    y = tf.squeeze(y)  #  Remove from the tensor shape a value of 1 Dimensions .
    y = tf.cast(y, dtype=tf.int32)
    y = tf.one_hot(y, depth=10)
    return x, y


batch_size = 128
(x, y), (x_val, y_val) = datasets.cifar10.load_data()
print('datasets:', x.shape, y.shape, x.min(), y.min())

train_db = tf.data.Dataset.from_tensor_slices((x, y))
train_db = train_db.map(pre_process).shuffle(1000).batch(batch_size)

test_db = tf.data.Dataset.from_tensor_slices((x_val, y_val))
test_db = test_db.map(pre_process).shuffle(1000).batch(batch_size)

sample = next(iter(train_db))
print('batch:', sample[0].shape, sample[1].shape)

network = MyNetwork()  # MYNetwork yes Keras.Model A subclass of 
network.compile(
    optimizer=optimizers.Adam(learning_rate=1e-3),
    loss=tf.losses.CategoricalCrossentropy(from_logits=True),
    metrics=['accuracy']
)
network.fit(train_db, epochs=50, validation_data=test_db, validation_freq=1)
network.evaluate(test_db)
network.save_weights('./ckpt/cifar10_weights.ckpt') # b Save the model to a disk file 

Reference link ：

1. Li Mu great God 《 Hands on deep learning 》TensorFlow Realization ,GitHub link ：https://github.com/TrickyGo/Dive-into-DL-TensorFlow2.0, Refer to CNN5.9GoogleNet Part of the code

2 Long Liangqu . Deep learning and TensorFlow Introduction of actual combat , project GitHub link ：https://github.com/dragen1860/Deep-Learning-with-TensorFlow-book, From which Lesson40--CIFAR And VGG actual combat