/h/ai_slop
TouchFluffyTails Them/They https://x.com/AyakonArts :marseydab2:  6mo ago (image post) 80 thread views #266898

Comparison on quality vs quantity for data in training a character

Unable to load image

:marseyneat:

12
New Old Top Bottom Controversial Random
Jump in the discussion.

No email address required.

So how's this work? You trained a LORA purely on prompted images? I've heard that's a big nono, but these turned out really well.

Or is this not an original character? Is this from an animu? I don't know animu. Is the original on the right something you prompted originally or something someone drew?

Jump in the discussion.

No email address required.

:#marseynietzsche:

Jump in the discussion.

No email address required.

Link copied to clipboard
Action successful!
Error, please refresh the page and try again.