模型简介: 使用此模型: · CFG scale: 3 · Lora强度:0.7 · 强烈建议使用highres.fix到 1.5x,使用以下选项之一: · · 升级器:无或任何你喜欢的降噪器:0.4(更接近原始但细节较少) · · 升级器:任何具有去噪功能的潜在升级器:0.5 或更高(更多细节,但略有变化) 该模型的主要触发词是: · cvttz06 car 您可以在肯定或否定提示中输入这些次要词语,以增加获得您喜欢的汽车图像的机会: · complete car · cropped · zoom · front view · rear view · 3/4 front left view · 3/4 front right view · 3/4 rear left view · 3/4 rear right view · side left view · side right view · zoom on headlamp · zoom on rear lamp · zoom on front wheel · zoom on rear wheel 很想看看您的图片,请上传! ~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~ 测试提示词和参数: Prompt: a closeup professional photo of cvttz06 car, zoom on headlamps, driving in the the city, raining , droplets, wet floor, water splashes, fine details, golden hour, hdr, 8k, ultra high-res, uhd, movie picture still, bokeh, sublime, breathtaking, glints, cropped <lora:cvttz06_car_t12_SDXL-000010:0.7> Negative prompt: complete car Steps: 40, VAE: sdxl_vae.safetensors, Size: 1024x768, Seed: 2812001709, Model: sd_xl_base_1.0, Version: v1.6.0, Sampler: DPM++ 2M Karras, VAE hash: 63aeecb90f, CFG scale: 3, Model hash: 31e35c80fc, Hires upscale: 1.5, Hires upscaler: Latent (bicubic antialiased), Denoising strength: 0.65, "cvttz06_car_t12_SDXL-000010: bfc39c4f02c1" |
1人表示很赞