Batch process, # of output depends on what input, interact with the output

Are these feasible in the hugging face gradio?

  1. Being able to batch process a bunch of images?
  2. The number of outputs depends on some processing of the inputs e.g. if there are two bounding box images, then output two images
  3. Is it possible to interact with the output - let’s say I have keypoints on image, can I drag to shift the location of the keypoint?

If so, is there any exemplar spaces I could look into to learn how to deploy these?

Thank you!