Finally V2 🤭 It should work better and with less tag
For V2
Seem to work well with base tags. Also have given up on trying to train open mouth... Maybe in future is possible?
ALWAYS USE IN NEGATIVE, I RECOMMEND
-mouth+, nose+, border around face, Neck, Back of head, looking away, face not in view, Empty eyes, Bad eyes, goofy eyes
For V1
Additional support tag, are best mostly for if adding style lora
tape deck on face, four buttons on top of head, handle on head
Have to specify what angle you want him from in both positive and negative
May have to write angles to make head work properly such as
3/4 view tag
3/4 view, from side, side view
If is from behind and looking back use tag
head from side, looking back,
and negative
-head backwards, broken neck,
Head entirely from behind
head from behind, back of head, switch on back of head, looking away
and then in negative
-eyes, tape deck
I also try training for open mouth as in the show... Is not working well, in v2 I will maybe try train with trigger word. But those tags are
open mouth, open tape deck, tape deck on face, mouth is a tape deck, flap mouth
And in negative?? Again is not working but maybe you can try
-tongue, closed tape deck, closed mouth
Always put in negative
-nose+, mouth+, neck, ears, border around face, robot, tv head, screen
If like my models leave review and post pictures 🤗❤️Thank you!