This sample demonstrates how to do image classification with the following approaches:

1. Use Azure Custom Vision Service
2. Train with traditional machine learning algorithms such as SVM
3. Train with Keras and Tensorflow

• azure_custom_vision_score uses the model you trained in Azure Custom Vision service to do predictions.
• process_images resizes and normalizes input images.
• image_classification_traditional_ml trains the clasifier with traditional ML algorithms.
• image_classification_tensorflow uses Keras to the train the model.
• train_deploy_azure_ml demonstrates how to train the model remotely with Azure Machine Learning Service, and use Azure Machine Learning Service to deploy the model as a web service. You would need to refactor your code in Jupyter notebook to Python scripts -
  • To train remotely, refactor your code to a script like train.py
  • To deploy, implement a script score.py and put the dependencies in env.yml
• deploy_without_aml reveals what's actually happening behind the scene when you deploy a service from a model.