Remix.run Logo
bee_rider 18 hours ago

The models are already trained… is inference so costly? It’s just a dot product or a matvec or something, right?