NoobAI-XL Detailer - V-Pred1_V1
Related Keywords & Tags
Highlighted images
Recommended Prompts
1girl, (himeyamato:0.9), aoi sakura \(seak5545\), (hyonee:1.1), (eversoul:0.9), achyue, (dishwasher1910:0.8), momoko \momopoco\, (nardack:0.8), (torino aqua:0.4), masterpiece, amazing quality, best quality, absurdres, very aesthetic, newest, very awa, sharp, crisp, 8k
Recommended Negative Prompts
worst quality, low quality, quality, lowres, worst aesthetic, jpeg artifacts, scan artifacts, signature, bad id, bad twitter id, english commentary, bad hands, username, mutated hands, mammal, logo, anthro, furry, scanlines, blur, blurry, comic, (long body, long neck:1.3), extra digit, missing digit, old, simple background, (censored:1.3), stomach bulge, cleft of venus, bar censor, twins, splatter, plump, leotard, wet, yellow skin, expressionless, steam, extra ears, (blurry background, bokeh:1.5), hair intakes, futanari, mosaic censoring, pixelated, school uniform, forehead jewel, half-closed eyes, cum, red eyeshadow, out of focus
Recommended Parameters
samplers
steps
cfg
resolution
other models
Recommended Hires (High-Resolution) Parameters
denoising strength
Tips
Use descriptive prompts like describing face and clothes to avoid overly dark or black silhouettes in images.
For V0.5 V-Pred version, a weight of 1.5 improves the image quality.
For V1.0 V-Pred version merged with custom flat color lora, use a recommended weight of 1.
For V0.5, weight range is recommended between 1-2, within the full range of -6 to +6.
If unsure about V-Pred versions, refer to the checkpoint description for details.
Version Highlights
This took waay more effort then it was supposed to. Ended up with tweaking OneTrainer to work with v-pred SDXL (thanks to guy from Discord for guiding me). But here it is:
Detailer V1, V0.5 merged with custom synthetic undertrained minimalist style lora. Please read main description for more info on specifics. Designed to be used with weight 1.
This is simple detailer trained on model NoobAI-XL.
V-Pred versions work better on v-pred model. If you don't know what it means - read checkpoint description. If not explicitly stated - 99% it is eps prediction version.
V-pred related specifics:
V0.5 - just slap it on with weight 1.5 makes image better.
V1.0 - slightly alters image giving more details. Weird interaction that I found - it darkens dark. V-pred NOOB-AI is first SDXL model that I see that can make really dark imagery. ESPECIALLY with short prompts, like 1girl, dark, night. If you are feeling that that is not what you want - just be more descriptive. describe face, vlothes and you won't get those black on black silhouettes anymore. Btw that applies to the NOOBvp1 itself too. See example here: https://civitai.com/images/47951726
On the upside it can darken stuff that base model has issues with darkening, like some poses.
V1.0
Merged with custom flat color lora, recommended weight is 1. Slightly alters image, especially with no prompts like 1girl.
V0.5
Weight range -6 - +6
I recommend using 1-2.
Contributor
Model Details
Discussion
Please log in to leave a comment.