How to Test Saved Model Using Never Before Seen Non-Labeled Images


#1

Hello,

I am creating a supervised model based on ~2500 labeled images (2 attributes) using AutoML, with an 80/20 train and test split, using 20-30 epochs. Once I have tuned the model, I would like to then separately validate about 200 images for which I have no labels for, to see if I can score them as part of an assignment I am working on for school.

How do I load and test unseen non-labeled images for use with a saved custom AutoML model in DLS? Thanks.

-AR


#2

Hello,
To test unseen data in DLS you need to first deploy the model as a web app then you can test your images.
or you can run inference on the test data by creating a zip file with a csv just like your training data but in this case name your CSV file as “test.csv”

For deployment of the model just go to the deploy tab and deploy your model with a click.

Regards
Rajat


#3

Thank you - I will try once I train my model.