Does prodigy tool support annotation of multimodal dataset?
I have a dataset where each data point consists of image and text. Each data point can be assigned multiple labels. Looking forward to hearing from you. Thanks
Hi @shuklaham and welcome to the Prodigy forum
We don't have a built-in recipe for multimodal inputs, but you can easily put together a custom UI by combining
choice interfaces via
If you need a custom layout/style, you can also inject custom html and css, but for a simple combination of images and text
blocks should be enough.
Here you can find the (simplified) code for the textcat recipes: https://github.com/explosion/prodigy-recipes/tree/master/textcat which can be helpful while working on your custom version of it.
Let us know if you need any help on the way