TensorFlow.NET 實戰(zhàn)

第一部分 TensorFlow.NET API 入門
第 1 章 TensorFlow.NET 介紹 ................................................................................... 2
1.1 TensorFlow.NET 特性 .................................................................................................. 2
1.2 TensorFlow.NET 開源庫結(jié)構(gòu) ...................................................................................... 3
第 2 章 數(shù)據(jù)類型與張量詳解 ........................................................................................ 6
2.1 數(shù)據(jù)類型 ...................................................................................................................... 6
2.2 張量詳解 ...................................................................................................................... 7
2.3 常量與變量 .................................................................................................................. 8
2.4 字符串常見操作 .........................................................................................................11
2.5 基本張量操作 ............................................................................................................ 14
2.6 維度變換 .................................................................................................................... 19
2.7 合并分割 .................................................................................................................... 22
2.8 廣播機制 .................................................................................................................... 24
第 3 章 Eager Mode 詳解 ............................................................................................ 28
3.1 Eager Mode 說明 ........................................................................................................ 28
3.2 Eager Mode 比較 ........................................................................................................ 29
3.3 Eager Mode 數(shù)值運算 ................................................................................................ 31
3.4 Eager Mode 張量降維運算 ........................................................................................ 32
3.5 Eager Mode 矩陣運算 ................................................................................................ 35
3.6 print 與 tf.print 特性對比 ........................................................................................... 37
第 4 章 自動求導(dǎo)原理與應(yīng)用 ....................................................................................... 44
4.1 機器學(xué)習(xí)中的求導(dǎo) .................................................................................................... 44
4.2 簡單函數(shù)求導(dǎo) ............................................................................................................ 45
4.3 復(fù)雜函數(shù)求偏導(dǎo) ........................................................................................................ 46
第 5 章 線性回歸實操 ...................................................................................................... 48
5.1 線性回歸問題 ............................................................................................................ 48
5.2 TensorFlow 下的線性回歸 ........................................................................................ 50
5.3 C#和 Python 的性能比較 .......................................................................................... 54
第 6 章 MNIST 手寫字符分類邏輯回歸 ............................................................................ 56
6.1 經(jīng)典的 MNIST 手寫字符分類問題 .......................................................................... 56
6.2 邏輯回歸代碼實操 .................................................................................................... 63
第 7 章 tf.data 數(shù)據(jù)集創(chuàng)建與預(yù)處理 ................................................................................ 77
7.1 tf.data 介紹 ................................................................................................................. 77
7.2 tf.data 數(shù)據(jù)集創(chuàng)建 ..................................................................................................... 78
7.3 tf.data 數(shù)據(jù)預(yù)處理 ..................................................................................................... 81
7.4 tf.data 數(shù)據(jù)使用 ......................................................................................................... 89
第 8 章 深度神經(jīng)網(wǎng)絡(luò)實踐 ............................................................................................................ 91
8.1 深度神經(jīng)網(wǎng)絡(luò)介紹 .................................................................................................... 91
8.2 TensorFlow.NET 代碼實操 1：DNN with Eager ...................................................... 93
8.3 TensorFlow.NET Keras 模型搭建的 3 種方式 ........................................................ 105
8.4 TensorFlow.NET 代碼實操 2：DNN with Keras .....................................................116
第 9 章 AutoGraph 機制詳解 .............................................................................................. 131
9.1 AutoGraph 機制說明 ............................................................................................... 131
9.2 AutoGraph 機制原理 ............................................................................................... 144
9.3 AutoGraph 編碼規(guī)范 ............................................................................................... 146
第二部分 .NET Keras 簡明教程
第 10 章 Keras 簡要介紹 ................................................................................................... 149
10.1 Keras 特性 .............................................................................................................. 149
10.2 Keras 版本說明 ...................................................................................................... 150
第 11 章 模型與層 ................................................................................................................... 152
11.1 Keras 常用的模型與層 .......................................................................................... 152
11.2 自定義層 ................................................................................................................ 155
11.3 自定義模型 ............................................................................................................ 157
11.4 模型常用 API 概述 ................................................................................................ 160
第 12 章 Keras 常用 API 說明 ........................................................................................... 167
12.1 回調(diào)函數(shù) ................................................................................................................ 167
12.2 數(shù)據(jù)集預(yù)處理 ........................................................................................................ 169
12.3 優(yōu)化器 .................................................................................................................... 172
12.4 損失函數(shù) ................................................................................................................ 175
12.5 評估指標 ................................................................................................................ 180
第 13 章 Keras 搭建模型的 3 種方式 ............................................................................ 184
13.1 Sequential API 方式 ............................................................................................... 185
13.2 Functional API 方式 ............................................................................................... 18