AWS SageMaker Complete Course | PyTorch & Tensorflow in NLP

本课程是 AWS SageMaker 的完整指南，学生将学习如何构建、部署 SageMaker 模型，方法是将本地 docker 容器集成到 SageMaker 中。课程还将深入探讨如何在 AWS SageMaker 环境中引入您自己的算法。课程还将解释如何使用预先构建的优化 SageMaker 算法。
课程还将深入推动如何创建管道和工作流，以便可以自动重新训练和安排模型。
本课程将为您提供有关如何在 Keras 中为多类分类用例构建 Transformer 框架的公平想法。另一种使用 Bert 等预训练模型解决多类分类问题的方法。
深度学习模型后来封装在本地机器的 Docker 中，然后再推回到 AWS ECR 存储库。
本课程提供：
什么是 SageMaker 以及为什么需要它
SageMaker 机器学习生命周期
SageMaker 架构
SageMaker 培训技术：
将您自己的 docker 容器
从本地带到 SageMaker将您自己的算法从本地机器带到 SageMaker
SageMaker 预构建算法
SageMaker 管道开发
安排 SageMaker 培训笔记本
提供超过 5 小时的课程，可帮助初学者在 SageMaker 中脱颖而出，并且将精通在 SageMaker 中构建、训练和部署模型
Created by Prakhar Agarwal | Published 6/2021
Duration: 5h 28m | 8 sections | 41 lectures | Video: 1280×720, 44 KHz | 2.2 GB
Genre: eLearning | Language: English + Sub

Build deep learning model in Tensorflow/Keras & PyTorch. How to bring docker container&Algorithm from local to Sagemaker

What you’ll learn
What is SageMaker and Why it is required
SageMaker Architechure
Model Building using existing Docker Image in SageMaker
Model Building using existing algorithm in SageMaker
Model Building using SageMaker Pre-built algorithms
Model Building in Tensorflow/Keras
Model Building in Pytorch
How to deploy the models in SageMaker
How to make predictions from Endpoints
Create complete End-to End machine learning Pipeline Workflow
Real time example of NLP
How to schedule the SageMaker notebook for Retraining
How to Build ,deploy and schedule the Model
Show more
Show less

Requirements
Free or paid subscription to AWS is required. It may ask for Phone and/or Credit Card for verificationPython Basic knowledge
Description
This course is complete guide of AWS SageMaker wherein student will learn how to build, deploy SageMaker models by brining on-premises docker container and integrate it to SageMaker. Course will also do deep drive on how to bring your own algorithms in AWS SageMaker Environment. Course will also explain how to use pre-built optimized SageMaker Algorithm.
Course will also do deep drive how to create pipeline and workflow so model could be retrained and scheduled automatically.
This course will give you fair ideas of how to build Transformer framework in Keras for multi class classification use cases. Another way of solving multi class classification by using pre-trained model like Bert .
Both the Deep learning model later encapsulated in Docker in local machine and then later push back to AWS ECR repository.
This course offers:
What is SageMaker and why it is required
SageMaker Machine Learning lifecycle
SageMaker Architecture
SageMaker training techniques:
Bring your own docker container from on premise to SageMaker
Bring your own algorithms from local machine to SageMaker
SageMaker Pre built Algorithm
SageMaker Pipeline development
Schedule the SageMaker Training notebook
More than 5 hour course are provided which helps beginners to excel in SageMaker and will be well versed with build, train and deploy the models in SageMaker
\n
Who this course is for:Data Engineers or data scientistDevelopers who want to start a career in or wants to learn about the exciting domain of Data Science and Machine LearningBusiness Analysts who want to apply Data Science to solve business problemsLearn how to build train and deploy it in AWS cloud