Actually, you don't really need a dedicated model as long as you have the proper adapter like the projector model for vision inputs you just need another for ro

Actually, you don't really need a dedicated model as long as you have the proper adapter like the projector model for vision inputs you just need another for robotic outputs, after that it is just a matter of having the training data.