[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]

/vt/ - Virtual Youtubers

Search:


View post   

>> No.48935128 [View]
File: 1.96 MB, 1024x1536, catbox_qhte72.png [View same] [iqdb] [saucenao] [google]
48935128

>>48926803
>thick thighs
what a coincidence, I'm genning some too
NAI worked way better with these loras :https://files.catbox.moe/fg6716.png

some ok gens I didn't bother fixing up (based64v3, will try 66 since it just released) :
https://files.catbox.moe/ey9mkf.png
https://files.catbox.moe/m6ut81.png
https://files.catbox.moe/c08p8f.png
https://files.catbox.moe/3xu3c3.png
https://files.catbox.moe/fj2rns.png
https://files.catbox.moe/ftimma.png


>>48927060
that's mostly because of the lowres at which it's trained and at which it generates. It would need to do some extreme latent ""antialiasing"" to do better. Which explains why it's able to do hands just fine when locally inpainting at higher res. Same with eyes, mouths, etc

Navigation
View posts[+24][+48][+96]