3 ms·
Ah, this is where this comes from. There has been rumours flying around in Stable Diffusion / Flux circles that you would get much more realistic pictures when
by jsemrau 2y ago
Ah, this is where this comes from. There has been rumours flying around in Stable Diffusion / Flux circles that you would get much more realistic pictures when you include a photo id like IMG_0416.
- Sharlin 2y agoI don’t think it comes from these Youtube videos – Flickr and other photo upload services are a more likely source of training images with default file names.
- jsemrau 2y agoMaybe its a combination of both.
- Sharlin 2y agoIt seems exceedingly unlikely to me that frames from random YouTube videos would have been used to train image generation models. First off, they're difficult to extract and second, the quality of individual video frames is very low, especially if we're talking about 15 year old phone videos at what, 480p at the very best!
- jsemrau 2y agoYou are probably right. I approached it from a high-value dataset perspective but would agree that fuzzy frames probably don't help much.
- aydyn 2y agoIts not a rumor, you really do and you can try it out yourself. Unfortunately its very finnicky and you cant really leverage it to produce a realistic image of what you want since any further prompting seems to override it. Its like a ghost in the machine prompt.
- jsemrau 2y agoMakes you wonder if its possible with the right seed, scheduler, and prompt to complete recreate the original.