Dirty result image - Githubissues

yuval-alaluf / stylegan3-editing

Official Implementation of "Third Time's the Charm? Image and Video Editing with StyleGAN3" (AIM ECCVW 2022) https://arxiv.org/abs/2201.13433

https://yuval-alaluf.github.io/stylegan3-editing/

MIT License

660 stars 72 forks source link

Dirty result image #49

Closed shartoo closed 10 months ago

shartoo commented 1 year ago

Running run_on_batch in utils/inference_utils.py with hyper parameter n_iters=3 and restyle_pSp_ffhq, the rebuild images seem dirty compared with original image.What can i do to avoid such result?

original images
rebuild images

YKJamesMoriarty commented 1 month ago

hi,how did you overcome this problem,i have the same dirty result too,please give me some help~thanks!