胡新悦, 汪海蓉, 张会雄. 基于Keras模型的深度学习实验教学研究[J]. 实验科学与技术, 2022, 20(3): 137-140. DOI: 10.12179/1672-4550.20210031
引用本文: 胡新悦, 汪海蓉, 张会雄. 基于Keras模型的深度学习实验教学研究[J]. 实验科学与技术, 2022, 20(3): 137-140. DOI: 10.12179/1672-4550.20210031
HU Xinyue, WANG Hairong, ZHANG Huixiong. Research on Experimental Teaching of Deep Learning Based on Keras Model[J]. Experiment Science and Technology, 2022, 20(3): 137-140. DOI: 10.12179/1672-4550.20210031
Citation: HU Xinyue, WANG Hairong, ZHANG Huixiong. Research on Experimental Teaching of Deep Learning Based on Keras Model[J]. Experiment Science and Technology, 2022, 20(3): 137-140. DOI: 10.12179/1672-4550.20210031

基于Keras模型的深度学习实验教学研究

Research on Experimental Teaching of Deep Learning Based on Keras Model

  • 摘要: 深度学习模型正在各行各业中发挥着重要作用,为了便于学生快速获得深度学习框架的上手经验,以增加他们的信心和学习兴趣,开发了该教学实验。教学实验的主要内容是指导同学们利用Flask轻量级的Web应用框架以及Keras人工神经网络库在两节课时间内运行一个基于深度学习的网页端应用。由于安装环境以及框架较容易使用与理解,学生们短时间内完成一个深度学习项目,能让学生看到这个深度学习项目的可视化结果并且有了上手的体验,完成该实验后基于该项目展开深度学习课程的讲解,非常适用于高校基于深度学习综合设计课程的开展。

     

    Abstract: Deep learning models are playing an important role in all walks of life. This teaching experiment is developed to facilitate students to quickly gain hands-on experience of the deep learning framework and increase their confidence and interest in learning. The main content of this teaching experiment is to guide students to run a deep learning-based web-side application in two classes sessions using Flask’s lightweight web application framework and keras artificial neural network library. Because the installation environment and framework are easier to use and understand, students are able to complete a deep learning project in a short period of time which allows students to see the visual results of this deep learning project and have a hands-on experience. After completing the experiment, they can start to learn more about deep learning based on the project. This experimental teaching mode is very suitable for the development of integrated design curriculum based on deep learning in colleges and universities.

     

/

返回文章
返回