|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/build-and-deploy-a-ml-model-to-production-with-aws-and-react/
课程评论：没有评论
课程名称：使用AWS和React部署生产级机器学习模型 课程概述：本课程将教您如何使用AWS Sagemaker、AWS API Gateway、Lambda、React.js、Node.js、Express.js、MongoDB以及DigitalOcean构建一个安全、可扩展且稳定的企业级图像分类器。我们将首先使用最佳实践和设置IAM策略，创建一个安全的AWS环境。接着，我们将利用AWS自带的SageMaker Studio Notebooks，学习如何使用任意自定义数据集，对数据集进行探索性数据分析，使用Matplotlib、Seaborn、Pandas和Numpy等工具。获取数据集的深刻见解后，我们将设置AWS的超参数调整任务，并展示如何使用GPU实例加速训练，以及如何进行多GPU实例的训练。然后，我们将评估训练工作，并查看精准率、召回率和F1分数等指标。在评估后，我们将通过AWS API Gateway和Lambda函数将深度学习模型部署到AWS上，并使用Postman测试我们的API，检查是否能获得推理结果。完成这些后，我们将保护我们的端点并设置自动扩展以防止延迟问题。最后，我们将构建与AWS API连接的Web应用，并将其部署到DigitalOcean。
In this course we are going to use AWS Sagemaker, AWS API Gateway, Lambda, React.js, Node.js, Express.js MongoDB and DigitalOcean to create a secure, scalable, and robust production ready enterprise level image classifier. We will be using best practices and setting up IAM policies to first create a secure environment in AWS. Then we will be using AWS' built in SageMaker Studio Notebooks where I am going to show you guys how you can use any custom dataset you want. We will perfrom Exploratory data analysis on our dataset with Matplotlib, Seaborn, Pandas and Numpy. After getting insightful information about dataset we will set up our Hyperparameter Tuning Job in AWS where I will show you guys how to use GPU instances to speed up training and I will even show you guys how to use multi GPU instance training. We will then evaluate our training jobs, and look at some metrics such as Precision, Recall and F1 Score. Upon evaluation we will deploy our deep learning model on AWS with the help of AWS API Gateway and Lambda functions. We will then test our API with Postman, and see if we get inference results. After that is completed we will secure our endpoints and set up autoscaling to prevent latency issues. Finally we will build our web application which will have access to the AWS API. After that we will deploy our web application to DigitalOcean.
