stable-diffusion icon indicating copy to clipboard operation
stable-diffusion copied to clipboard
verified publisher icon CompVis

Evaluation issuse

Open nieshenx opened this issue 3 years ago • 1 comments

Thank you for your great work! But i have some question when calculating text-to-image fid. In the appendix <E.3.2 Text-to-Image Synthesis>, I notice that you said that fid is calculated by comparing generated samples with 30000 samples from the validation set of the MS-COCO dataset. But in the paper <Zero-Shot Text-to-Image Generation> which you followed, the fid is calculated by comparing 30000 generated samples with the validation set of the MS-COCO dataset.

nieshenx avatar Sep 29 '22 06:09 nieshenx

What is the difference?

LukeLIN-web avatar Jan 16 '25 17:01 LukeLIN-web