Any idea how this scores on the text-to-image generation benchmark COCO: https://paperswithcode.com/sota/text-to-image-generation-on-coco https://paperswithcod

Any idea how this scores on the text-to-image generation benchmark COCO: https://paperswithcode.com/sota/text-to-image-generation-on-coco https://paperswithcode.com/sota/text-to-image-generation-on-... Based on the few examples provided, appears much weaker than the state of the art.