V2 was designed to be much simpler, it doesn't have near as many tags to prompt, but what it has in return is much more consistency and accuracy in your prompts, (it still kind of sucks?). I'll keep V1 up because it has a much wider variety of tags but in terms on accuracy, consistency, it's much worse.
maybe combine them together?
"****" is the trigger tag
penis prompts for the bottom
More V2 Information
~0.8-1 LoRa strength
Cfg 5
Seems to be pretty stuck with an anime style, use your favorite realism checkpoint to cure this
Sketchy style tag was meant to be a negative prompt
V1 information
"****" and "Penis" seem to be mandatory tags to produce a good image
Check the tags, it's got some good facial expressions
trained on "feet" and "2boys" tag but they suck, try it anyway maybe it will do something good.
LoRa strength 0.95
start with cfg 7
adetailer used on some faces
