模型介绍: Stylized 3D Model LoRA This is a LoRA trained on stylized 3D model images. It can provide a figure-like material and a distinctive silhouette. Changes in 3.0 ▣ Increased the training resolution (ver.2: 768, ver.3: 896) ▣ Revised the tagging ▣ ▣ Removed the trigger word as well ▣ ▣ Reviewed the training parameters Improvements in 3.0 ▣ It can reproduce a more stereoscopic effect than ver.2 ▣ LoRA can be applied without a trigger word ▣ The effect of LoRA can be adjusted by adding the 3d tag to either Positive or Negative Prompt Precautions Depending on the base model, the face may become overly deformed. If it doesn't work well, consider using the following models, which are used for verification: ▣ Facebomb Mix ▣ Alstroemeria Mix ~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~ 示例提示词和参数: Prompt: <lora:stylized3DModelLora_v30:0.8>, hyper detailed 3d, 1girl, looking at viewer, big hair, absurdly long hair, sitting, full body, clutter girl's lovely room, hugging stuffed animal, fluffy hoodie with animal ears, masterpiece, best quality, 8k, detailed skin texture, detailed cloth texture, beautiful detailed face, intricate details, ultra detailed, zentangle, <lora:hipoly3DModelLora_v10:0.2:NCNF>, hiqcgbody, Negative prompt: EasyNegative, extra fingers,fewer fingers, zentangle, (simple background:1.5), floating object, makeup, face paint, mole on face, Steps: 20, CFG scale: 7, Sampler: DPM++ 2M Karras, Seed: 902579781, ENSD: 31337, Size: 512x896, Model: models_02_25D_facebombmix_v1, Model hash: aad629159b, Hires upscale: 1.5, Hires upscaler: Latent, Denoising strength: 0.6, Clip skip: 2 |