模型简介: 这个模型源于我个人经常使用的lora和ckpt的融合,介于2.5D和3D之间,个人审美偏好的结果。以下是模型的混合公式: 0.9(0.75(LuckyStrikeMixv0.1)+0.25(chilloutmix_Ni-fp16)+(koreanDollLikeness_v15:0.2,shojovibe_v11:0.3,fashionGirl_v45:0.4,LoraKAigirlFace_kAigirlFace2:0.5,eyeLora_eyesV10:0.6:OUTD))+0.1(chilloutmix_Ni-fp16) 这个融合的过程中在2.5D和3D之间摇摆不定,最终效果只是我个人的选择。VAE 已在此模型中烘焙,因此无需加载任何额外的 VAE。感谢以上作者提供的优秀模型。 顺便一提,所有的预览图像都是使用这个模型生成的,没有使用额外 LORA。 ~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~ 主图提示词和参数: Prompt: masterpiece, best quality, 1girl, black hair, blush, butterfly, hair ornament, dress, bare feet, flower, hair ornament, medium breasts, leaf, plant, puddle, purple eyes, short hair, squatting, toes, water, wet clothes, blue fashion dress , cloud, cloudy sky, building, blue sky, depth of field, Negative prompt: easynegative, badhandv4, nsfw Steps: 25, Size: 512x768, Seed: 1130040547, Model: Breeze-Mix_v3fp16, Sampler: DPM++ 2M Karras, CFG scale: 7, Clip skip: 2, Model hash: 1b76c4c015, Hires steps: 20, Hires upscale: 2, Hires upscaler: Latent, Denoising strength: 0.55 |