Possibly that would apply to text generation, but not specific for image classification or segmentation tasks.

Comments