Author: Zen and the Art of Computer Programming
1 Introduction
Since 1990, deep learning has received more and more attention. With the increase in data volume, computing performance, and machine learning model complexity, the field of deep learning is gradually becoming more and more popular. There are currently many deep learning frameworks, such as TensorFlow, Caffe, Torch, PaddlePaddle, etc. The most popular deep learning framework in recent years is Keras, open sourced by Google.
In June 2017, Keras officially released version 1.0, which is the first stable version and supports Python 2.7 and Python 3.6+. It has the characteristics of easy to get started, rapid development, strong scalability, and model modularity. In March 2019, Keras underwent an important upgrade - version 2.3.1, which added support for TensorFlow 2.x. In this version, Keras has reconstructed the underlying implementation, reducing the difference between TensorFlow 1.x version and 2.x version, making Keras more in line with Python concepts.
In this article, I will introduce the basic use of Keras through some code examples, from entry to proficiency. I hope it can help you get started quickly and master the skills of using Keras.
The main content of this article includes: Keras overview, Keras installation and environment construction, Keras basic model construction, Keras data loading and preprocessing, Keras training and verification, Keras saving and loading, Keras migration learning, Keras integrated learning, etc. .
2.Keras Overview
Keras is a high-level neural network API whose goal is to make simple things simple and complex things feasible. It can run on Theano or TensorFlow and supports GPU computing and deep learning practices. This article will be based on the Keras v2.3.1 version.
Keras mainly consists of the following aspects:
Sequential Model : It is a linear sequence model,