/vt/ - Virtual Youtubers

Batch 2
Bonnie Barskwell (globie) PonyXL LoRA https://mega.nz/folder/dfBQDDBJ#3RLMrU3gZmO6uj167o-YZg/file/YCgEWAib
Kumanui Miel (globie) PonyXL LoRA https://mega.nz/folder/dfBQDDBJ#3RLMrU3gZmO6uj167o-YZg/file/5Xw2CCQY
CookieSwirlC PonyXL LoRA https://mega.nz/folder/dfBQDDBJ#3RLMrU3gZmO6uj167o-YZg/file/JGwEkaYS
Hanatan PonyXL LoRA https://mega.nz/folder/dfBQDDBJ#3RLMrU3gZmO6uj167o-YZg/file/Qf4GxCZR

All loras made with very limited datasets. Settings result in decently close but not fully accurate clothing, accurate faces/hair, and probably restrictive styles (I haven't been mixing styles a lot because the style loras are all huge and annoying to switch around), results surprisingly flexible (e.g https://files.catbox.moe/anywiv.png )

Batch 3 will probably be Reina Ronronea, Taring Hu, Tomoe Carmine... Might take some suggestions, but my current settings seem to work best with datasets like this as more diverse datasets seem to result in more underbaked loras (though I haven't fully experimented with this, like throwing out art that veers a bit too far for comfort)