AI爱好者
上一篇文章中介绍了如何通过LoRA训练个性风格 AI绘画：手把手训练风格化LoRA教程，LoRA是相对低成本的对模型微调的手段，如果能够直接微调模型，效果必然是更加出色的。
今天介绍的迪士尼皮克斯动画风，就是网友最新训练分享的微调模型（非LoRA），用于绘制皮克斯风格的图画。
用法
Stable Diffusion加载底模可以看之前的系列基础教程： 无需显卡的 AI 绘图：手把手 Colab 部署 Stable Diffusion（上）｜ 无需显卡的 AI 绘图：手把手 Colab 部署 Stable Diffusion（下）。
加载完成后直接出图当然不在话下：
masterpiece, high quality best quality,1girl, bangs, beach, blue_sky, blush, bow, breasts, checkered, checkered_shirt, checkered_skirt, cloud, cloudy_sky, collarbone, day, envelope, giving, grass, hair_bow, heart, holding, holding_letter, horizon, incoming_gift, kazami_yuuka, leaning_forward, lens_flare, letter, light_rays, long_hair, looking_at_viewer, love_letter, mountain, mountainous_horizon, ocean, outdoors, plaid, plaid_background, plaid_bikini, plaid_bow, plaid_bowtie, plaid_bra, plaid_dress, plaid_headwear, plaid_jacket, plaid_legwear, plaid_necktie, plaid_neckwear, plaid_panties, plaid_pants, plaid_ribbon, plaid_scarf, plaid_shirt, plaid_skirt, plaid_vest, pov, shirt, skirt, sky, smile, solo, sun, sunbeam, sunlight, tree, unmoving_patternNegative prompt: EasyNegative, drawn by bad-artist, sketch by bad-artist-anime, (bad_prompt:0.8), (artist name, signature, watermark:1.4), (ugly:1.2), (worst quality, poor details:1.4), bad-hands-5, badhandv4, blurrySize: 512×768, Seed: 170953175, Model: Disney Pixar Cartoon type A, Steps: 20, Sampler: Euler a, CFG scale: 7.5, Clip skip: 2, Model hash: d6548414b4, Hires steps: 20, Hires upscale: 2, Hires upscaler: Latent, Denoising strength: 0.5
masterpiece, high quality best quality,1girl, animal_ears, kimono, bangs, blue_eyes, blurry, blurry_background, blush, breasts, cleavage, hair_between_eyes, large_breasts, lips, long_hair, looking_at_viewer, smile, solo, upper_body, window
Negative prompt: EasyNegative, drawn by bad-artist, sketch by bad-artist-anime, (bad_prompt:0.8), (artist name, signature, watermark:1.4), (ugly:1.2), (worst quality, poor details:1.4), bad-hands-5, badhandv4, blurry
Size: 512×768, Seed: 1746483825, Model: Disney Pixar Cartoon type A, Steps: 20, Sampler: Euler a, CFG scale: 7.5, Clip skip: 2, Model hash: d6548414b4, Hires steps: 20, Hires upscale: 2, Hires upscaler: Latent, Denoising strength: 0.5
更可以把现有素材转换为迪士尼风格：
实现方式是在Stable Diffusion的img2img中将原图输入，并在 Controlnet 组件中将原图作为 OpenPose 素材输入即可。
欢迎免费使用GPT对话，感受ChatGPT的魅力！AI爱好者 – 最具实力的中文AI交流社区平台 (aiahz.com)
长按扫描二维码进群领资源
