Source: Deep Learning on Medium
How to Inference a Pre-trained Image Classifier Using Watson AutoAI
Inference AutoAI pre-trained models
IBM Watson Studio – AutoAI
AutoAI is an IBM platform that allows faster and easier management for the AI lifecycle pipeline. It provides better algorithms and techniques to understand your data, clean the data, train, fine-tune, and deploy different models in few minutes. It also allows using pre-trained models. In this post, we will use a pre-trained image classifier to inference different images.
Image classification is another task that falls under Data Science and Machine Learning where we assign each image one or more classes or categories.
Create a project on IBM Watson. You should be able to see a list of your projects. Click on one of the projects to start.
Click on “Add to project” to add a service
Choose “Visual Recognition Model”
If it is your first time, then you will be redirected to create a service. Click “here”
You may choose one of your previous services (if you have any) or create a new one.
Choose a plan that fits your needs.
You have the option to change the region, plan, resource group, etc. I prefer the defaults.
After that, you should be able to see three different trained models that you can use.
Let us start with “General” . After clicking “General”, you will see three tabs. Overview will display the model info.
Click “Test” to start inference the model
Drag and drop the images you want to inference. Each image will be displayed with its associated classes along with their confidence score.
Click “Implementation” to be provided with few ways to inference your model remotely using a terminal.
Let us try the “Food” model.
“Explicit” model predictions