I've cracked the code to generate consistent mouthless images with my LoRA! As such, this will be my guide to make a baseline image in Stable Diffusion for mouthless women! Positive Prompt: a (mouthless:1.4) ((mouthless)) woman without [a mouth and lips:0.25], mouthless ((((celeb)))) in a really nice color summer halter top and a color skirt in front a beautiful mountaintop, ((mouthless:1.25)), (silent), (a mouthless woman:1.4), mouthless, mouthless, mouthless, Mouthless woman, no mouth, (speak no evil), [upper lip:0.1], [lower lip:0.1], [mouth:0.1], clothing, wearing clothing, mouthless woman with clothing, woman who can speak no evil, silent, mouthless gag, gagged by flesh, skin gag, dressed in clothing, <lora:Mouthless_V1.3:1.4>, <lora:Mouthless:1.3>, Negative Prompt: paintings, sketches, worst quality, low quality, normal quality, lowres, normal quality, monochrome, grayscale, skin spots, acnes, skin blemishes, age spot, glan, extra fingers, fewer fingers, strange fingers, bad hand, bad anatomy, fused fingers, missing leg, mutated hand, malformed limbs, cross eyed, ugly faces, distorted faces, looking cross-eyed, weird eyes, ((mouth, lips, lipstick, teeth, smile, unhappy, lips:1.5)), mouth, lips, lipstick, teeth, smile, unhappy, lips, upper lip, lower lip, mouth, lips, lipstick, teeth, smile, unhappy, lips, upper lip, lower lip, mouth, lips, lipstick, teeth, smile, unhappy, lips, upper lip, lower lip, nudity, nsfw, naked, risque, tits out, nipples,
I also want to put you through which checkpoints I used when making it. They are placed from best (on top) to worst (on the bottom), but it might require fine tuning on your end!
Realism By Stable Yogi (https://civitai.com/models/166609?modelVersionId=325825)
AnalogMadness (https://civitai.com/models/8030/analog-madness-realistic-model)
DreamShaper 8 (https://civitai.com/models/4384?modelVersionId=128713) The Truality Engine (https://civitai.com/models/158621?modelVersionId=288888)
Project AiO (https://civitai.com/models/18428?modelVersionId=245967)
EpicRealism Pure Evolution / Natural Sin (https://civitai.com/models/25694?modelVersionId=143906)
RealDream (https://civitai.com/models/153568?modelVersionId=270190)
Source of the Goddess (https://civitai.com/models/241450?modelVersionId=344233) It's great on the mouth, but the weaker backgrounds hold it back a ton! A lot of potential
EpicPhotogasm LastUnicorn (Has some issues with color flatness) (https://civitai.com/models/132632?modelVersionId=223670)
Epicdream_Lullaby (https://civitai.com/models/125189?modelVersionId=136739)
ICBINP - "I Can't Believe It's Not Photography" (New Year edition) (https://civitai.com/models/28059?modelVersionId=306898)
Final Reality 1.0 (https://civitai.com/models/162177?modelVersionId=182619)
RealisticVision V6 (https://civitai.com/models/4201?modelVersionId=130072) just does weird shit.
Cyberrealistic v4.1 Back to Basics (wildly inconsistent) (https://civitai.com/models/15003?modelVersionId=256915)
Link to wildcards and LoRA's: https://drive.google.com/drive/folders/1YUdwvaGzBbWPRC7IS3VAiZKxzMUHvq3A?usp=sharing
I ran every test in the Euler a sampler with 48 sampling steps. I set the CFG Scale between 2-3. I set the image size to a width of 512 and the height to 768 before upscaling (this can be flipped for landscape images) without loss of quality. Upscaling I applied the R-ESRGAN 4x+ with a denoising step between 0.60 and 0.90 and upscaled by a factor of 1.5 (even though higher will work too, keep in mind to lower the denoising if needed). I found 1.5 to be more consistent though. The most important thing is to fiddle with it and see what works best. Hopefully this has given you information to generate your own images though!
Here are some images!