Thank you very much for your work! May I ask if you have trained the model on the object detection data with only the object box as the prompt