How to increase the resolution or clarity of the result in simswap and increase the characteristic of the source in the output video (in single face swap)

[TreblaSOS]

Respected Developer and friends... i am new to this i have done all the thing that have mention in the guide... and it works in simswap... but the resulting video is not to my expectation. could you pliz tell me what additional steps must be done to make the source face more visible and realistic in the output video in simple face swap. I use this command in Swap only one face within the video(the one with highest confidence by face detection) https://github.com/neuralchen/SimSwap/blob/main/docs/guidance/usage.md

the result video is ( i used iron man image, but i can see only little resemblance of iron man.. how to increase the source characteristics more on the resulting video.)

most of the time... the swaped face in result video is somewhat blurry or low resolution as compare to race of the background. how to make it natural increase its compatibility with the back ground or the race of the body in the video.

and lastly how can i produce high quality result like the one demonstrated in the main page of simswap-

Sorry for the trouble... and thank you

[Fibonacci134]

Hey, have you tried adding the 512 parameter? It's the one based on the high-quality face sets.

[SrGrafoasspee]

how to use 512 parameters?

[szopeno]

quite simple: test_wholeimage_swapsingle.py --crop_size 512

don't forget to download 512.zip, just as it is explained in the docs.