@Daggoth  That makes sense! =8) from a consumer point of view; I want her at 18 years old, less make-up, with braces, in a neat summer dress. See? That's what consumers do. They'll tweak your work in ways you didn't think of. Just a thought ;) Again, thanks

@Daggoth  Yes, wise choice. Always try to include a special "trigger word" for each

individual model.

Many people do not know, but you DO NOT want your trained images for

your model to be mixed in with DATA/Pictures which are already inside your SD-1.5 model

program. Because sometimes some of those images inside the program are of a star/celebrity

which may look like 50 years old. There were 1996 famous faces (and pornstars) which are

inside the original program.

Also try to use at least 15-to-25 pictures of a face from different angles, AND use images ONLY

from a certain age and not mix ages. Example ; pictures of a star from 21 years old ONLY and

not mix pictures from them being 40 or 50 years old. This I learned from making my own Lora

amd T.I. models for myself.

DO Train them on the base model for SD-1.5 because they will be better trained on it and then

can be used with ANY other available checkpoints. The faces will always be better and consistent.