3 ms·
I feel like I'm looking at aliens wearing people's skin. https://kota.is/qpxa1.png https://kota.is/qpxa1.png Nightmare inducing. I hate it. https://kota.is/X57
by brink 4y ago
I feel like I'm looking at aliens wearing people's skin. https://kota.is/qpxa1.png https://kota.is/qpxa1.png
Nightmare inducing. I hate it. https://kota.is/X57J5.png https://kota.is/X57J5.png
Ugh, this legit makes my stomach turn. https://kota.is/Q2hUI.jpg https://kota.is/Q2hUI.jpg
It's like a bad amusement park ride; no thanks.
- flangola7 4y agoThis is worse than SD 1
- notahacker 4y agoNot sure what's scarier, the faces or the message from Chtulu in the text! Unless these examples were cherry picked rare fails, I'm surprised it's so bad when there are a few of other generative models out there including ones by OpenAI which do human faces almost perfectly. Unless they're really throttling processing power used?
- barking_biscuit 4y agoThe models at the moment don't do well on scenes where the faces are small and far away. I assume it's due to training on 256px by 256px images and not being able to capture enough detail in that, so the gens look horrible like this. At present, when using Stable Diffusion, you can address this as part of your workflow by fixing the faces with inpainting / upscaling. This can also be automated in a scriptable way using the Unprompted extension.
- kossTKR 4y agoWow this is not production ready! What is microsoft doing? Look at midjourney 5's absolute perfection as an example.