f0442166cd
* Updated curated environments in sample notebooks * Fixed continuous retraining notebook
ONNX on Azure Machine Learning
These tutorials show how to create and deploy Open Neural Network eXchange (ONNX) models in Azure Machine Learning environments using ONNX Runtime for inference. Once deployed as a web service, you can ping the model with your own set of images to be analyzed!
Tutorials
- If you are using an Azure Machine Learning Notebook VM, you are all set. Otherwise, Configure your Azure Machine Learning Workspace
Obtain pretrained models from the ONNX Model Zoo and deploy with ONNX Runtime
- MNIST - Handwritten Digit Classification with ONNX Runtime
- Emotion FER+ - Facial Expression Recognition with ONNX Runtime
Train model on Azure ML, convert to ONNX, and deploy with ONNX Runtime
Demo Notebooks from Microsoft Ignite 2018
Note that the following notebooks do not have evaluation sections for the models since they were deployed as part of a live demo. You can find the respective pre-processing and post-processing code linked from the ONNX Model Zoo Github pages (ResNet, TinyYoloV2), or experiment with the ONNX models by running them in the browser.
- ResNet50 - Image Recognition with ONNX Runtime
- TinyYoloV2 - Convert from CoreML and deploy with ONNX Runtime
Documentation
Related Articles
- Building and Deploying ONNX Runtime Models
- Azure AI – Making AI Real for Business
- What’s new in Azure Machine Learning
License
Copyright (c) Microsoft Corporation. All rights reserved.
Licensed under the MIT License.
Acknowledgements
These tutorials were developed by Vinitra Swamy and Prasanth Pulavarthi of the Microsoft AI Frameworks team and adapted for presentation at Microsoft Ignite 2018.
