zlacker

[return to "RealFill: Image completion using diffusion models"]
1. derefr+P01[view] [source] 2023-09-30 01:57:53
>>flavor+(OP)
Question: does this model only do outpainting, or does it also do super-resolution? Could this model be fed all the frames of a really awful security-camera video, in order to then synthesize a high-resolution still image of a suspect?
◧◩
2. pbhjpb+Ll1[view] [source] 2023-09-30 07:54:06
>>derefr+P01
And you can tell it in advance which person you want it to reveal as the suspect!
[go to top]