对图像进行坐标检测，生成的bbox是resize成正方形之后的值吗？ about mplug-owl

[fansticOne]

请问如何设计prompt来对图像进行坐标检测，输出bbox？感谢

[zhaop-l]

请问如何设计prompt来对图像进行坐标检测，输出bbox？谢谢

Please provide the bounding box coordinate of the region this sentence describes : people on car.

[fansticOne]

请问如何设计prompt来对图像进行坐标检测，输出bbox？谢谢

Please provide the bounding box coordinate of the region this sentence describes : people on car.
试了没有给出bbox呀，回答的是：
The bounding box coordinate of the region this sentence describes is the sheep lying on the ground.

[LukeForeverYoung]

Yes, the images are resized to squares for example 448x448. But, the generated coordinates should be a value in the range [0,1], which are ratios that are unrelated to the actual resolution. Therefore, whatever you resize the image, the coordinates are always applicable. You may need to check if the image has been cropped, as this could cause the coordinates to be offset relative to the original image.

[fansticOne]

请问一下，对图像进行坐标检测，使用的是哪个权重文件？