I trained a tiny 14M param model that outperforms some big boys on one task: Selecting a model to accomplish a task!

I trained a tiny 14M param model that outperforms some big boys on one task: Selecting a model to accomplish a task! Model here: https://huggingface.co/pmarquees/succinct-router Write up here: https://www.pedromarques.io/succinct-router/