Author: Zen and the Art of Computer Programming
1 Introduction
1.1 Why write this article?
Keras is an open source deep learning library written in Python and has an advanced, flexible and friendly interface. Keras can help developers more conveniently implement the process of building, training, and optimizing neural network models, thereby greatly improving the efficiency of deep learning development. However, as a deep learning framework, its own internal mechanism is not yet easy to understand. Therefore, this article hopes to lead you to quickly get started with Keras through an introductory tutorial and find ways to solve practical problems. At the same time, we will combine different types of problems to show readers how to use Keras for various aspects of deep learning in the form of examples. Finally, we will also try to answer some questions readers may have and share some Keras-related resources for your reference. This article can not only help novices learn the correct use of Keras, but also provide intermediate and senior engineers with an in-depth understanding of Keras information and provide guidance and reference for deep learning projects.
1.2 Author introduction
The author currently works at Tencent AI Lab, focusing on the research and development of deep learning systems. He has been engaged in research and development in the fields of image processing, machine vision, and speech recognition for many years, and has conducted in-depth research on machine learning algorithms in computer vision, natural language processing and other fields. Nowadays, his research direction is biased towards computer vision, and he has been committed to exploring how to improve human life through intelligent methods.
1.3 Directory structure
- Chapter 1Basic knowledge of model construction and training
- 1.1 TensorFlow and Keras
- 1.2 Basic process of model construction
- 1.3 Basic principles of model training process
- 1.4 Data set partitioning, batch size, and data enhancement methods
- 1.5 Activation function and regularization method
- 1.6 Compilation parameter settings
- 1.7 Model saving and adding