深度学习——学习笔记（1）神经网络基础

# -*- coding: utf-8 -*- 
# @Time : 2020/12/25 16:30 
# @Author : Renlele 
# @File : 2_1.py 
# @Software: Pycharm
def func1():
    # 加载数据集
    from keras.datasets import mnist
    (train_images, train_labels), (test_images, test_labels) = mnist.load_data()
    print(train_images.shape)  # (60000, 28, 28)
    print(train_labels)  # [5 0 4 ... 5 6 8]
    # 查看数据类型
    print(train_images.dtype)  # uint8
    print(test_images.shape)  # (10000, 28, 28)
    print(test_labels)  # [7 2 1 ... 4 5 6]

    # 使用matplotlib库显示数字
    import matplotlib.pyplot as plt
    digit = train_images[4]
    plt.imshow(digit,cmap = plt.cm.binary)
    plt.show()


    from keras import models
    from keras import layers

    network = models.Sequential()
    # Dense层  全连接
    network.add(layers.Dense(512, activation='relu', input_shape=(28 * 28,)))
    network.add(layers.Dense(10, activation='softmax'))

    # 编译
    # categorical_crossentropy 损失函数，学习权重张量的反馈信号，应使他最小化
    # rmsprop 优化器
    network.compile(optimizer='rmsprop',
                    loss='categorical_crossentropy',
                    metrics=['accuracy'])

    # 准备图像数据
    train_images = train_images.reshape((60000, 28 * 28))
    train_images = train_images.astype('float32') / 255

    test_images = test_images.reshape((10000, 28 * 28))
    test_images = test_images.astype('float32') / 255

    # 准备标签
    from keras.utils import to_categorical
    train_labels = to_categorical(train_labels)
    test_labels = to_categorical(test_labels)

    # 调用fit方法完成在训练数据上的拟合
    network.fit(train_images, train_labels, epochs=5, batch_size=128)

    # output
    # Epoch 1/5
    # 469/469 [==============================] - 4s 8ms/step - loss: 0.2580 - accuracy: 0.9250
    # Epoch 2/5
    # 469/469 [==============================] - 3s 7ms/step - loss: 0.1039 - accuracy: 0.9697
    # Epoch 3/5
    # 469/469 [==============================] - 3s 7ms/step - loss: 0.0688 - accuracy: 0.9794
    # Epoch 4/5
    # 469/469 [==============================] - 4s 8ms/step - loss: 0.0497 - accuracy: 0.9855
    # Epoch 5/5
    # 469/469 [==============================] - 4s 8ms/step - loss: 0.0380 - accuracy: 0.9887

    # 检查在测试集上的性能
    test_loss, test_acc = network.evaluate(test_images, test_labels)
    print('test_acc:', test_acc)


def func2():
    import numpy as np
    # 0D张量
    x = np.array(12)
    print(x)  # 12
    print(x.ndim)  # 0

    # 1D张量  == 向量
    x = np.array([12, 3, 6, 14, 7])
    print(x)  # [12  3  6 14  7]
    print(x.ndim)  # 1

    # 矩阵(2D张量）  行与列
    x = np.array([[5, 78, 2, 34, 0],
                  [6, 79, 3, 35, 1],
                  [7, 80, 4, 36, 2]])
    print(x)  # [[ 5 78  2 34  0] [ 6 79  3 35  1] [ 7 80  4 36  2]]
    print(x.ndim)  # 2

    # 3D张量
    x = np.array([[[5, 78, 2, 34, 0],
                   [6, 79, 3, 35, 1],
                   [7, 80, 4, 36, 2]],
                  [[5, 78, 2, 34, 0],
                   [6, 79, 3, 35, 1],
                   [7, 80, 4, 36, 2]],
                  [[5, 78, 2, 34, 0],
                   [6, 79, 3, 35, 1],
                   [7, 80, 4, 36, 2]]
                  ])
    print(x.ndim)  # 3