Cryosibilum Style | Illustrious - Experimental
Prompts recommandés
mcht2, 1girl, detailed image, (masterpiece:1.8), masterfully drawn, detailed background, (best quality:1.6), expressive, (ultra detailed:1.8), (high quality rendering:1.9), best resolution, very awa, (8k:1.6), latest, amazing quality, 4k, very aesthetic, <lora:3728b893-9cf0-4025-a0b5-ddbee4ef5169:0.9>, absurdres, intricate details, incredibly absurdres
Prompts négatifs recommandés
3d, ai-generated, deleted, lowres, chibi, loli, lolita, child, children, kid, petite, worst aesthetic, (bad anatomy, bad hands:1.4), abstract, bad proportions, distorted, fisheye, bad feet, bad perspective, bad fingers, extra fingers, fused fingers, missing fingers, extra arms, extra legs, duplicate limbs, extra limbs, missing limbs, mutated hand, malformed limbs, misaligned eyes, asymmetrical eyes, cross-eyed, cross eyed, wrong hand, wrong foot, cloned, repetitive, off-topic, jpeg artifacts, (signature:1.4), twitter username, patreon username, logo, (watermark:1.5), (artist name), (censor:1.8), (bar_censor:1.9), text, disembodied penis, side cutout, dark penis, dark testicles, (lactation), lactating, milk, cum_on_face, facial, cum_on_body, cum_on_hair, excessive_cum, motion_lines, speed_lines, selfie, reaching
Paramètres recommandés
samplers
steps
cfg
clip skip
resolution
vae
other models
Paramètres haute résolution recommandés
upscaler
upscale
steps
denoising strength
Conseils
Use a LoRA strength of 0.9; lower values give more flexibility.
Keep CFG scale below 4.5 for best results.
Avoid using too many LoRAs together to prevent rapid quality degradation.
High CFG values can diminish image quality.
Points forts de la version
This is still being tested, experimental version. I'll be reworking the dataset after the feedback I get from this release + I've been testing it a bit myself aswell, I know it is a bit undertrained and not so flexible as I want it to be.
Sponsors du créateur
If you want to support the creator, consider donating, liking, or following for more updates. More support options are available here.
Hey there, so I'll be releasing this one after testing it for a bit, I'm baffled at the high quality I've gotten so far. This LoRA was specifically trained on Illustrious v2.0 (the stable version) and I don't know if it works with other Illustrious models out there, so I'll have to test that aswell. I don't know if I even need to rework the dataset or not, the quality is amazing already, so I'll be waiting for feedback on this one... hopefully you guys enjoy it aswell. If you want to support me, you can either donate, like or follow for more updates, either one helps! Thank you!
Recommended settings:
Strength 0.9, lower for more flexibility. Same goes for your cfg, anything below 4.5 is perfect or recommended. Also try not to use too many LoRAs together as it can deminish the quality of your image fairly quickly, or a high cfg will also do that.
Thank you again for using my LoRA, you can support me through either donating Buzz, or you give a thumbs up, follow me or you can also use one of the links provided beneath my profile, anything helps really.
Contributeur
Détails du modèle
Type de modèle
Modèle de base
Version du modèle
Hash du modèle
Mots entraînés
Créateur
Discussion
Veuillez vous log in pour laisser un commentaire.