ONNX runtime to serve AI models

Invalid data source. Please correct the following errors:
  • The specified Sheet Name (Workshops (2)) was not found. Please try again with the correct one from the following list:\nWindows Server 2019 Microsoft Azure 2 React Microsoft 365 Projektų valdymas Asmeninis efektyvumas Scrum ITIL4F ITIL4P ITIL4L

Course overview:

The workshop will cover the basics of a Machine Learning project, from start to production release. We will concentrate on the optimization part of the chosen model. You will learn to use the ONNX Runtime to serve the model, check the performance compared to the initial model, and use a programming language different from the starting one.

Objectives

The participants will learn how to use the ONNX Runtime to optimize and serve their AI models.

Target audience

The workshop is open to all those who want to know more about Machine Learning and how an AI model goes from the study to the production phase. The first parts are introductory and will make everyone adept at creating their test model.

Technical requirements

– a Google Colaboratory (free) account
– a Google Drive (free) account
– basic notions of programming, Python

Duration: 1 day

Agenda:

Part 1: Setup

  • Colab
  • Connect Drive and
  • Libraries

Part 2: Study our model

  • Clean our dataset
  • Choose the algorithms
  •  Study the model
  • Save the model

 Part 3: Introduction to ONNX

  • The AI open format
  • Export the model in ONNX
  • Use Netron

Part 4: ONNX Runtime

  • Load the ONNX model
  • Play with the options
  • Serve the model
  • Benchmark

Part 5: Laboratory

  • Create model
  • Export ONNX model
  • Serve with ONNX runtime
  • Validate the process