Resized to 83% of original (view original)
Prompt | 1girl, solo, multicolored hair, ((gradient hair), white+(gold)+(blue:0.5) hair//), very long hair, messy+dishevled hair, two colored hair, streaked hair, bangs, 1fang, fangs, :o, Open Mouth With Cute Fangs Out, masterpiece, best quality, Thick Layered Medium Twintail Hair, Blush outside:1.4, in the background magestic+fantastic ancient cityscape, (Highest Quality, Amazing Details:1.4), Masterpiece, Bloom, Picturesque, <lora:outfit-sword_maiden-pynoise:0.9> outfit-bladenun, blindfold:1.1, habit, blindfolded, black blindfold, wet, rain in the background, walking:1.4 outside in the rain, view from above:1.2, <lora:VisNoctisLGHT:0.6> |
---|---|
Negative prompt | , bad-artist bad-artist-anime bad-hands-5 bad_prompt bad_prompt_version2 easynegative EasyNegativeV2 verybadimagenegative_v1.3 FastNegativeV2 Asian-Less-Neg, (worst quality, low quality:1.4) |
Sampler | DPM++ 2M Karras |
Seed | 956360138 |
Steps | 51 |
Cfg Scale | 7 |
Model Hash | 82376d9409 |
Noise offset lora comparison:
VisNoctisLGHT (lora block adjusted [style] VisNoctis 15) vs LowRA
VisNoctisLGHT is better at closeup shots. You can get original and block adjusted here:
https://huggingface.co/Aotsuyu/LoRA/blob/main/artstyles/%5Bstyle%5D%20VisNoctis%2015.safetensors
https://huggingface.co/Jemnite/ALunarLight/blob/main/Lua's%20Light/materials/VisNoctisLGHT.safetensors
#AIart #AIイラスト #stablediffusion #餃子Mix