How to train an image classification model with 0 code?

Azure Custom Vision is just for that

1/ Setting up

With an active Azure subscription, create a Custom Vision Resource

After that create a Custom Vision Project in the Custom Vision Portal

In our case, we are going to classify pet animals
2/ Adding images

Upload the images of the pet animals

Label the images, in our case as Cat, Dog, and Rabbit.

Proceed to quick train the model for our classifier
3/ Training the model

The model training should take a few minutes

When the model iteration was done in the previous step it should provide the Precision, Recall, and AP performance metrics

These should all be high to have a model with high precision
4/ Test the model

Above the performance metrics, click on the quick test

Upload another image of a cat apart from the trained images

The probability for cat should be the highest

Proceed to publish the model
5/ Configure and run

Now come back to the Azure portal

Now you can use the model as an application by running it in the cloud shell

Simple, without any code, with the power of Azure you can build such an Image classification model!
That's a wrap!

If you enjoyed this thread:

1. Follow me @freest_man for more of these
2. RT the tweet below to share this thread with your audience

Recommended by
Recommendations from around the web and our community.

Great share! Azure's ML workspace is also worth checking out! Lets you do everything from data labelling, to training and deployment in a seamless manner.