目 录
1 numpy 基础知识 ..............................................................................................................................................4
2 卷积 .....................................................................................................................................................................4
2.1 卷积的基本概念 .................................................................................................................................4
2.2 图像卷积 ............................................................................................................................................ 10
3 各种 GAN 变体 ............................................................................................................................................. 13
3.1 生成式模型 ........................................................................................................................................ 13
3.1.1 GAN .......................................................................................................................................... 13
3.1.2 DCGAN .................................................................................................................................... 14
3.2.3 各种 GAN .............................................................................................................................. 15
3.2.4 InfoGAN .................................................................................................................................. 16
3.2.5 AC-GAN .................................................................................................................................. 16
4 CNN 与 TensorFlow 实现 ........................................................................................................................... 16
4.1 CNN 的引入 ....................................................................................................................................... 16
4.2 CNN 的层次结构 .............................................................................................................................. 16
4.2.1 输入层 .................................................................................................................................... 17
4.2.2 卷积层 .................................................................................................................................... 17
4.2.3 激励层 .................................................................................................................................... 20
4.2.4 池化层 .................................................................................................................................... 21
4.2.5 归一化层 ................................................................................................................................ 22
4.2.6 切分层 .................................................................................................................................... 23
4.2.7 融合层 .................................................................................................................................... 23
4.2.8 全连接层和输出层 ............................................................................................................. 24
4.3 典型 CNN ........................................................................................................................................... 24
4.3.1 LeNet-5 ................................................................................................................................... 24
4.3.2 AlexNet.................................................................................................................................... 25
4.3.3 VGGNet ................................................................................................................................... 26
4.3.4 GoogleNet ............................................................................................................................. 27
4.3.5 ResNet 模型 ........................................................................................................................... 33
4.4 CNN 实现 ............................................................................................................................................ 33
4.4.1 参数及结构 ........................................................................................................................... 33
4.4.2 实现代码 ................................................................................................................................ 34
4.4.3 代码解析 ................................................................................................................................ 36
5 TensorFlow 入门学习 .................................................................................................................................. 42
5.1 鸢尾花 ................................................................................................................................................. 42
5.2 特征列 ................................................................................................................................................. 48
5.3 数据集 ................................................................................................................................................. 49
5.4 检查点 ................................................................................................................................................. 50
5.5 TensorBoard ....................................................................................................................................... 50
5.5.1 介绍 ......................................................................................................................................... 50
5.5.2 体验 ......................................................................................................................................... 51
5.5.3 探究 ......................................................................................................................................... 55
5.6 TensorFlow 学习例子 ...................................................................................................................... 60
