Pixtral Capabilities Regarding Input Bounding Boxes
#35
by maurovitale - opened
Does Pixtral support the use of input bounding boxes for tasks such as grounded captioning? If so, what is the correct method for passing in input bounding boxes?
Does Pixtral support the use of input bounding boxes for tasks such as grounded captioning? If so, what is the correct method for passing in input bounding boxes?