1、tensorflow版本查看

import tensorflow as tfprint('Tensorflow Version:{}'.format(tf.__version__))
print(tf.config.list_physical_devices())

2、MNIST数据集下载与预处理

(train_images,train_labels),(test_images,test_labels) = tf.keras.datasets.mnist.load_data()
train_images,test_images = tf.cast(train_images/255.0,tf.float32),tf.cast(test_images/255.0,tf.float32) # 归一化

tf.data.Dataset制作训练数据集

ds_train_image = tf.data.Dataset.from_tensor_slices(train_images)
ds_train_label = tf.data.Dataset.from_tensor_slices(train_labels)
ds_train = tf.data.Dataset.zip((ds_train_image,ds_train_label))
ds_train = ds_train.shuffle(10000).repeat().batch(64) # 乱序，无限次重复，每次取64张图片print(ds_train_image)
print(ds_train_label)
print(ds_train)

tf.data.Dataset制作测试数据集

ds_test = tf.data.Dataset.from_tensor_slices((test_images,test_labels))
ds_test = ds_test.repeat().batch(64)print(ds_test)

3、搭建MLP模型

from keras import Sequential
from keras.layers import Flatten,Dense
from keras import Inputmodel = Sequential()
model.add(Input(shape=(28,28)))
model.add(Flatten())
model.add(Dense(units=256,kernel_initializer='normal',activation='relu'))
model.add(Dense(units=10,kernel_initializer='normal',activation='softmax'))
# model.summary()

4、模型训练

调用model.compile()函数对训练模型进行设置

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

调用model.fit()配置训练参数，开始训练，并保存训练结果。

steps_per_epochs = train_images.shape[0]//64  # 937H = model.fit(ds_train,  # 训练数据集steps_per_epoch=steps_per_epochs,  # 每个epoch训练步数validation_data=ds_test,  #验证数据集validation_steps=10000//64,epochs=10,verbose=1)