MIT License
Microsoft
In this workshop, you will build an image classifier with Azure Custom Vision Service and export it as a TensorFlow model. You will learn how to import the model into a Python application and use it to classify images.
Follow the following instructions to create a Custom Vision project. You can use the training images linked to in the tutorial to build a cats vs dogs image classifier.
Tutorial: Train and export a TensorFlow model with Azure Custom Vision Service
If you are feeling adventurous or if you have another idea in mind, you can download and use your own images to train a model that recognizes other objects.
If you haven't cloned this repo already, clone it to your local machine.
In a terminal, browse to the app
folder.
Create and activate Python 3 virtual environment by entering:
python3.7 -m venv .venv
source .venv/bin/activate
# windows:
# .venv\scripts\activate
Copy the model.pb
and labels.txt
that you downloaded from Azure Custom Vision Service into model folder or use the ones that are already there.
Run the script and a prediction should appear:
python run.py
Azure Custom Vision Service can export models that run on many platforms.
Also check out these other resources: