[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/vt/ - Virtual Youtubers

Search:


View post   

>> No.72123410 [View]
File: 146 KB, 1238x772, no_lora_ceres fauna_by aztodio_00005_.jpg [View same] [iqdb] [saucenao] [google]
72123410

>>72122435
Fuck! Maybe there is another method, I'll look it up.
Proof I'm not talking out of my ass this time. This is AniXL3.1 with 0.25 unet lyco extracted from pony, 0.0 clip and no other loras.
Chuubas still work, styles still work, score prompting is not required.

https://files.catbox.moe/t9cziq.png
https://files.catbox.moe/k822mj.png
https://files.catbox.moe/7l6o5d.png
https://files.catbox.moe/snvfkr.png
https://files.catbox.moe/uvka1w.png

If I can determine the best extraction settings and merge it into the model without text encoder that might work as well but needs more testing. (If anyone with comfy wants the extracted lyco just ask.)

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