Practical Deep Learning on the Cloud

深度学习和机器学习应用程序正在成为技术和传统公司中许多业务的支柱。一旦组织在使用 ML/AI 算法方面取得首次成功，他们经常面临的主要问题是如何自动化和扩展其 ML/AI 工作流程。本课程将帮助您在云上更快地设计、开发和训练深度学习应用程序，而无需花费过多的时间和金钱。

本课程将大量利用当代公共云服务，例如 AWS Lambda、Step 函数、Batch 和 Fargate。无服务器基础设施可以大规模并行处理数千个请求。您将学习如何解决 ML 和数据工程师在以经济高效的方式训练许多模型以及构建数据管道以实现高可扩展性时遇到的问题。我们介绍了一些涉及使用预先训练的卷积神经网络模型来解决计算机视觉任务的技术。你将建立一个深度学习训练管道；解决多框架、并行训练和成本优化等问题；并通过导入预训练的卷积神经网络模型并将其用于您的项目来节省时间。

MP4 | Video: h264, 1280×720 | Audio: AAC, 44.1 KHz, 2 Ch
March 27, 2020 | ISBN: 9781838820374 | English
Duration: 27 Lessons (2h 27m) | Size: 1.38 GB

Build deep learning applications from scratch and deploy them on the cloud in a simple and cost-effective way

Learn
Training, exporting, and deploying deep learning models on the cloud (TensorFlow)
Using pre-trained models for your computer vision task
Working with cluster infrastructures on AWS (AWS Batch and Fargate)
Creating deep learning pipeline for training models using AWS Batch
Creating deep learning pipelines to deploy a model into production with AWS Lambda and AWS Step functions
Creating a data pipeline using AWS Fargate

About
Deep learning and machine learning applications are becoming the backbone of many businesses in both technological and traditional companies. Once organizations have achieved their first success in using ML/AI algorithms, the main issue they often face is how to automate and scale up their ML/AI workflows. This course will help you to design, develop, and train deep learning applications faster on the cloud without spending undue time and money.

This course will heavily utilize contemporary public cloud services such as AWS Lambda, Step functions, Batch and Fargate. Serverless infrastructures can process thousands of requests in parallel at scale. You will learn how to solve problems that ML and data engineers encounter when training many models in a cost-effective way and building data pipelines to enable high scalability. We walk through some techniques that involve using pre-trained convolutional neural network models to solve computer vision tasks. You’ll make a deep learning training pipeline; address issues such as multiple frameworks, parallel training, and cost optimization; and save time by importing a pre-trained convolutional neural network model and using it for your project.

By the end of the course, you’ll be able to build scalable and maintainable production-ready deep learning applications directly on the cloud.

The code files and related files are uploaded on GitHub at https://github.com/PacktPublishing/-Practical-Deep-Learning-on-the-Cloud

Features
Easily train and deploy scalable deep learning models on the cloud
Master AWS services while working with computer vision tasks and neural networks
Automate and scale your workflow with limited resources to gain maximum efficiency