Develop Custom Object Detection Models with NVIDIA and Azure Machine Learning

Overview

Module 1: Create workspace resources for getting started with Azure Machine Learning.
In this module, you'll learn how to:
- Create an Azure Storage account for storing and loading image data
- Create workspace resources for getting started with Azure Machine Learning
  - Create an Azure Machine Learning Studio workspace
  - Create an Azure Machine Learning Studio compute instance for use in model in training / validation
  - Create an Azure Machine Learning Studio Datastore
Module 2: Create a labeled dataset using Azure Machine Learning data labeling tools.
In this module, you'll learn how to:
- Create a Dataset of labeled images retrieved from an attached Datastore
- Coordinate data, labels, and team members to efficiently manage and process labeling tasks
- Track progress and maintain a queue of incomplete labeling tasks
- Review labeled data and export as an Azure Machine Learning Dataset
Module 3: Use AutoML to train a labeled dataset and develop a production model.
In this module, you'll learn how to:
- Author AutoML models for vision tasks via the Azure ML Python SDK
- Seamlessly integrate with the Azure Machine Learning data labeling capability
- Optimize model performance by specifying the model algorithm and tuning the hyperparameters
- Operationalize at scale, leveraging Azure Machine Learning MLOps capabilities
- Download the resulting model for use in a production deployment
Module 4: Deploy model to NVIDIA Triton Inference Server.
In this module, you'll learn how to:
- Create an NVIDIA GPU Accelerated Virtual Machine
- Configure NVIDIA Triton Inference Server and related prerequisites
- Execute an inference workload on NVIDIA Triton Inference Server

Syllabus

Module 1: Module 1: Create workspace resources for getting started with Azure Machine Learning
- Introduction
- Introduction to Azure Machine Learning
- Create an Azure Storage Account
- Create an Azure Storage Container
- Create an Azure Machine Learning Workspace
- Create an Azure Machine Learning Compute Instance
- Create an Azure Machine Learning Datastore
- Knowledge check
- Summary
Module 2: Module 2: Create a labeled dataset using Azure Machine Learning data labeling tools
- Introduction
- Create an Azure Machine Learning data labeling project
- Label images with Azure Machine Learning data labeling tools
- Export a labeled Azure Machine Learning dataset
- Knowledge check
- Summary
Module 3: Module 3: Use AutoML to train a labeled dataset and develop a production model
- Introduction
- Prepare the Jupyter notebook workspace
- Configure the Jupyter notebook execution environment
- Execute the Jupyter Notebook to produce an object detection model using AutoML
- Knowledge check
- Summary
Module 4: Module 4: Deploy model to NVIDIA Triton Inference Server
- Introduction
- Create a GPU Accelerated Virtual Machine
- Install prerequisites and NVIDIA Triton Inference Server
- Execute inference workload on NVIDIA Triton Inference Server
- Knowledge check
- Summary