Machine learning models with ONNX and .NET

Most machine learning models are trained in Python using frameworks like PyTorch and TensorFlow. You might expect that an application that runs the trained model would need to have all those frameworks installed. But does it need to? It does not! By converting your model to ONNX, you can run it in any supported language, including your .NET applications! ONNX (Open Neural Network Exchange) is an open standard for machine learning interoperability. It enables tools to work together by allowing them to share models in a commonly understood format. In this presentation we will show you how you can train a machine learning model using transfer learning in PyTorch, we’ll convert the model to ONNX, and we’ll use it in a .NET application.

Transcript

1. Machine learning models with ONNX and .NET Beatriz Stollnitz Principal Cloud Advocate @ Microsoft Henk Boelman Senior Cloud Advocate @ Microsoft
2. What are we building? Lego Dataset Homer Simpson 99.9% ?
3. Before we start
4. Simulation of human intelligence by machines. Artificial Intelligence
5. Machine Learning Ability to learn without being explicitly programmed.
6. Programming Algorithm Data Answers
7. Machine Learning Algorithm Data Answers
8. Machine Learning Model Data Answers
9. Machine Learning Model Data Answers
10. Machine Learning Predictions Data Model Data Answers
11. Deep Learning Subset of machine learning that's based on artificial neural networks.
12. What is Computer Vision? Computer vision is a field of artificial intelligence that trains computers to interpret and understand the visual world.
13. In transfer learning, a model trained for one task is used as a starting point for a different task. What is Transfer Learning?
14. PyTorch is an open-source deep learning framework originally developed by Meta AI, and now part of the Linux Foundation. What is PyTorch?
15. ONNX is an open format built to represent machine learning models. It defines a common set of operators and a common file format to enable AI developers to use models with a variety of frameworks, tools, runtimes, and compilers. What is ONNX?
16. What are we building? Lego Dataset Homer Simpson 99.9% ?
17. What is Azure Machine Learning? Set of Azure Cloud Services Python SDK AML Studio or Interface of your choice That enables you to:  Prepare Data  Manage Models  Build Models  Track Experiments  Train Models  Deploy Models
18. Azure ML service Key Elements Workspace Models Environments Experiments Deployment Pipelines Datastores Compute Target Data Labelling
19. Interface
20. ML Pipelines An Azure Machine Learning pipeline is an independently executable workflow of a complete machine learning task. Subtasks are encapsulated as a series of steps within the pipeline. An Azure Machine Learning pipeline can be as simple as one that calls a Python script, so may do just about anything. Workflows of steps that can use Data Sources, Datasets and run on Compute targets
21. Create a step / component Input Output Runs a script on a Compute Target in a Docker container. Parameters
22. Create a pipeline Dataset of Simpsons Images Train the Model with a PyTorch model Register the model Blob Storage Account Model Management model Convert Model To ONNX
23. Azure ML Designer What is the Azure ML Designer? AML Designer is an UI interface which enables you to build machine learning pipelines with drag-n-drop experience and simplify the publishment and deployment of your pipelines.
24. Demo Building an Azure ML Pipeline and run an ONNX model in .NET MAUI
25. What did we build? Lego Dataset Homer Simpson 99.9%
26. Thank you Aka.ms/ai/onnx-maui