Tensor.Art
Create

More real face, with bit lips and wink eyes

LORA
Reprint


Updated:

1K

Creator's description on Civitai:

Hey Guys and Girls ... (iam in FLUX contest)

works only with close portraits

trigger:

(biting own lips) , (winks with one eye at viewer) , (artistic makeup with ... )

I think this LoRa has a good potential for much more real facial expressions especially at lower weights ~0.4, also eyes very good, have fun ! may add "angry" or "sad" etc (mostly more natural teeth)

FLUX - only closeup works fine, start from my prompts, (art makeup dont work)

v53 - bit stronger (only 1 of 10 are good), on low weight very good eyes and facial expression and the slightly asymmetrical face (use negative prompt for makeup or, winks etc so its all in one lora and SDXL has problems to seperate)

v52 - should work for most models including my own

https://civitai.com/models/270312?modelVersionId=304720

v1 - for jugger - https://civitai.com/models/133005?modelVersionId=166909

v2 - for yamer - https://civitai.com/models/127923?modelVersionId=170252

v3 - not better but more compatible to other models (if you copy the prompt, i renamed the lora)

v4 - for demon core - https://civitai.com/models/155977?modelVersionId=177633

watch my prompts and play with it, weight between 0.4 and 1, CFG arround 5, steps at least 40 (works only with portraits)

besides the artistic makeup, the facial expressions are much more real, even if only 1 out of 10 pictures does exactly what it is supposed to do, e.g. bite the lips or wink the eyes.

at lower weights <0.8, the mouth is sometimes slightly slanted or one eye is half closed or both are half closed, which is much more realistic than the expression of a model baby's face.

here is another faceexpression blow kiss

https://civitai.com/models/287540/blow-a-kiss-by-dice

.

.

have fun ...

Version Detail

FLUX.1
1000
20
https://civitai.com/user/ElectronLibre My goal with this lora was to achieve a typical biting lip expression with a large defined hump formed by the lower lip. It is trained on some close ups, I selected an epoch that showed a great balance in flexibility and concept adherence, but you should put key words about eyes, hair and anything you want to see in your final image. Aside from that, it is bias towards forward facing, viewed from the front picture because this expression doesn’t mean much from the side, but it should have great flexibility for the body angle and position.

Project Permissions

Model reprinted from : https://civitai.com/models/151594?modelVersionId=754028

Reprinted models are for communication and learning purposes only, not for commercial use. Original authors can contact us to transfer the models through our Discord channel --- #claim-models.

Comments

Related Posts

Describe the image you want to generate, then press Enter to send.