Discovering Bugs in Vision Models using Off-the-shelf Image Generation and Captioning

Open in new window