This is awesome. I’m guessing the models are executed on the database server and not a separate cluster? What about GPU training? How is that handled? I’d love

This is awesome. I’m guessing the models are executed on the database server and not a separate cluster? What about GPU training? How is that handled? I’d love to see more docs.