keras多层感知器识别手写数字

一、进行数据预处理

步骤1：导入所需模块

import tensorflow as tf
import tensorflow.keras
import numpy as np
import pandas as pd
from tensorflow.keras.utils import to_categorical
np.random.seed(10)

步骤2：读取MNIST数据

from tensorflow.keras.datasets import mnist
(x_train_image, y_train_label), \
(x_test_image, y_test_label) = mnist.load_data()

步骤3：将features（数字图像特征值）以reshape转换

x_train=x_train_image.reshape(60000,784).astype('float32')
x_test=x_test_image.reshape(10000,784).astype('float32')

步骤4：将features（数字图像特征值）标准化

x_train_normalize = x_train/255
x_test_normalize = x_test/255

步骤5：label（数字真实的值）以One-Hot Encoding进行转换

y_trainonehot =tf.keras.utils. to_categorical(y_train_label)
y_testonehot = tf.keras.utils.to_categorical(y_test_label)

二、建立模型
1.导入所需模块

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense

2.建立Sequential模型
建立“输出层”和“隐藏层”

model = Sequential()
model.add(Dense(units=256,
               input_dim=784,
               kernel_initializer='normal',
               activation='relu'))
model.add(Dense(units=10,
               kernel_initializer='normal',
               activation='softmax'))

3.查看模型的摘要

print(model.summary())

三、进行训练

1.定义训练方式

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

2.开始训练

train_history=model.fit(x=x_train_normalize,
                       y=y_trainonehot,
                       validation_split=0.2