没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
试读
575页
Better Deep Learning Train Faster, Reduce Overfitting, and Make Better Predictions by Jason Brownlee 26 step-by-step lessons, 575 pages. quotes from papers and books. step-by-step tutorial projects. 深度学习神经网络已经变得易于定义和拟合,但仍难以配置。 在这部以您习惯的友好机器学习掌握风格编写的新电子书中,准确了解如何提高深度学习神经网络模型在预测建模项目上的性能。 通过清晰的解释、标准的 Python 库(Keras和TensorFlow 2)以及分步教程课程,您将发现如何更好地训练模型、减少过度拟合和做出更准确的预测。
资源推荐
资源详情
资源评论
Better Deep Learning
Train Faster, Reduce Overfitting,
and Make Better Predictions
Jason Brownlee
i
Disclaimer
The information contained within this eBook is strictly for educational purposes. If you wish to apply
ideas contained in this eBook, you are taking full responsibility for your actions.
The author has made every effort to ensure the accuracy of the information within this book was
correct at time of publication. The author does not assume and hereby disclaims any liability to any
party for any loss, damage, or disruption caused by errors or omissions, whether such errors or
omissions result from accident, negligence, or any other cause.
No part of this eBook may be reproduced or transmitted in any form or by any means, electronic or
mechanical, recording or by any information storage and retrieval system, without written permission
from the author.
Acknowledgements
Special thanks to my proofreader Sarah Martin and my technical editors Andrei Cheremskoy, Michael
Sanderson, Arun Koshy.
Copyright
Better Deep Learning
©
Copyright 2019 Jason Brownlee. All Rights Reserved.
Edition: v1.3
Contents
Copyright i
Contents ii
Preface iii
Introduction v
Welcome v
Framework for Better Deep Learning x
Diagnostic Learning Curves xix
I Better Learning 1
1 Improve Learning by Understanding Optimization 3
1.1 Neural Nets Learn a Mapping Function . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Learning Network Weights Is Hard . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Key Features of the Error Surface . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Navigating the Non-Convex Error Surface . . . . . . . . . . . . . . . . . . . . . 8
1.5 Implications for Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Components of the Learning Algorithm . . . . . . . . . . . . . . . . . . . . . . . 9
1.7 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Configure Capacity with Nodes and Layers 14
2.1 Neural Network Model Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Nodes and Layers Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Model Capacity Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
ii
CONTENTS iii
3 Configure Gradient Precision with Batch Size 29
3.1 Batch Size and Gradient Descent . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2 Gradient Descent Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.3 Batch Size Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4 Configure What to Optimize with Loss Functions 49
4.1 Loss Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Regression Loss Functions Case Study . . . . . . . . . . . . . . . . . . . . . . . 56
4.3 Binary Classification Loss Functions Case Study . . . . . . . . . . . . . . . . . . 64
4.4 Multiclass Classification Loss Functions Case Study . . . . . . . . . . . . . . . . 74
4.5 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.6 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5 Configure Speed of Learning with Learning Rate 87
5.1 Learning Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2 Learning Rate Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.3 Learning Rate Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6 Stabilize Learning with Data Scaling 121
6.1 Data Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.2 Data Scaling scikit-learn API . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.3 Data Scaling Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
7 Fix Vanishing Gradients with ReLU 141
7.1 Vanishing Gradients and ReLU . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
7.2 ReLU Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.3 ReLU Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
7.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
8 Fix Exploding Gradients with Gradient Clipping 167
8.1 Exploding Gradients and Clipping . . . . . . . . . . . . . . . . . . . . . . . . . . 167
8.2 Gradient Clipping Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
8.3 Gradient Clipping Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
8.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
8.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
CONTENTS iv
8.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
9 Accelerate Learning with Batch Normalization 180
9.1 Batch Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
9.2 Batch Normalization Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
9.3 Batch Normalization Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . 189
9.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
9.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
9.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
10 Deeper Models with Greedy Layer-Wise Pretraining 201
10.1 Greedy Layer-Wise Pretraining . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
10.2 Greedy Layer-Wise Pretraining Case Study . . . . . . . . . . . . . . . . . . . . . 204
10.3 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
10.4 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
10.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
11 Jump-Start Training with Transfer Learning 221
11.1 Transfer Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
11.2 Transfer Learning Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
11.3 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
11.4 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
11.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
II Better Generalization 243
12 Fix Overfitting with Regularization 245
12.1 Problem of Model Generalization and Overfitting . . . . . . . . . . . . . . . . . 245
12.2 Reduce Overfitting by Constraining Complexity . . . . . . . . . . . . . . . . . . 247
12.3 Regularization Methods for Neural Networks . . . . . . . . . . . . . . . . . . . . 248
12.4 Regularization Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . 249
12.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
12.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
13 Penalize Large Weights with Weight Regularization 252
13.1 Weight Regularization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
13.2 Weight Regularization Keras API . . . . . . . . . . . . . . . . . . . . . . . . . . 258
13.3 Weight Regularization Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . 260
13.4 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
13.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
13.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
14 Sparse Representations with Activity Regularization 272
14.1 Activity Regularization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
14.2 Activity Regularization Keras API . . . . . . . . . . . . . . . . . . . . . . . . . 276
14.3 Activity Regularization Case Study . . . . . . . . . . . . . . . . . . . . . . . . . 278
剩余574页未读,继续阅读
ThinkSpatial空间思维
- 粉丝: 1183
- 资源: 64
下载权益
C知道特权
VIP文章
课程特权
开通VIP
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功
- 1
- 2
前往页