Results of Evaluation #14
Description
Hi, thanks for your great work!
I try to generate images using your provided checkpoints and run sample_ADE20K.sh. However, the generated images seem a little different from your provided one here. I got 26.2 FID and 37.0 mIoU, which are different from those reported in the paper.
If i understand right, the results should be consistent under the fixed checkpoints and seeds. Is the provided model the best one used in your paper, or did you use a different setting during image generation (e.g., random seed, batch size, sampler)?