Introducing the Orthogonal Model of Emoji

Based on my Orthogonal Model of Emotions, I’ve created a derivative dataset that removes one dimension and replaces the ontological cluster labels with 10 emoji. You know, as one does! I then validated the dataset with Create ML using a Transfer Learning sequence and 15 training iterations.

Once the dataset is uploaded, I’ll create a HF model and space as a demo. Please check out my other work here on HF and let me know what you think! For instance, do you want to create a derivative set with equally weighted categories? What else can you think of to try?

Made in with