Collection of VLMs model fine-tuned to predict a single action end-effector location/position to reach a target based on the prompts and camera feeds