第二章 神经网络的数学基础

下面展示一些 内联代码片。

# -*- coding: utf-8 -*-
""" 
手写数字分类-MNIST数据集 
"""
from keras.datasets import mnist    # 加载Keras中的MNIST数据集

(train_images, train_labels), (test_images, test_labels) = mnist.load_data()

print(train_images.shape)   # train_images 3D
print(len(train_images))
print(train_labels)
print(train_images.ndim)   # 轴的个数
print(train_images.dtype)  # 数据类型


# 1.网络架构
from keras import models
from keras import layers

network = models.Sequential()     # 调用Sequential()方法(构造器)
network.add(layers.Dense(512, activation='relu', input_shape=(28*28,)))   # 全连接层
network.add(layers.Dense(10, activation='softmax'))                       # 多分类
network.summary()

# 2.预处理
network.compile(optimizer='rmsprop',                # rmsprop优化器
                loss='categorical_crossentropy',    # 损失函数：categorical_crossentropy：分类交叉熵
                metrics=['accuracy'])               # 指标：精度

# 3.准备图像数据
train_images = train_images.reshape((60000, 28*28))   # 变成2D  形状(60000, 784)
train_images = train_images.astype('float32') / 255   # 预处理，取值范围0～1
test_images = test_images.reshape(10000, 28*28)
test_images = test_images.astype('float32') / 255

# 4.准备标签：将数据列表转换为整数张量(标签)
from keras.utils import to_categorical

train_labels = to_categorical(train_labels)
test_labels = to_categorical(test_labels)

network.fit(train_images, train_labels, epochs=5, batch_size=128)      # 用训练集训练模型
test_loss, test_acc = network.evaluate(test_images, test_labels)       # 评估测试集
print('test_loss:', test_loss)
print('test_acc:', test_acc)

代码运行结果: