← 返回排行榜

Full Fine Body Moss Image by Sarcastic TOFU 姿势/镜头

sarcastictofu · · MOSS Image

姿势/镜头类 LoRA,作者 sarcastictofu,底模为 MOSS Image,触发词 Full_Fine_Body。

特点

适用范围

关键参数

适配底模
MOSS Image
触发词
Full_Fine_Body
累计下载
19
发布

本机部署建议(RTX 5060 Ti 16GB)

16GB 结论
LoRA 本身只占几百 MB,显存够;瓶颈在底模。

使用建议

作者原文摘录

This LORA is designed to generate outputs with a diverse range of full female body shot (head to feet, head to toes) poses. The LORA is trained on a good dataset of high quality images featuring models in various poses, captured from different angles and under varying lighting conditions showcasing the entire body in Indoor or Outdoor settings. This LORA can create very high quality AI Image outputs. Simply use these keywords to gain proper aesthetics for your subjects with this LORA, however even without it if you have good prompt it can give you good outputs- Full_Fine_Body also optionally you can add these secondary keywords - front view shot - to specify a front view full body shot bac…

作者例图请在 Civitai 模型页 查看(站外,含成人内容风险,不做内嵌)。

官方 ComfyUI 工作流

下载 JSON 后拖进 ComfyUI 窗口即可载入;缺节点通常是 ComfyUI 版本太旧。

官方链接

下载 / 安装

文件级链接点一下就直接下载(HF 会 302 到 CDN);每条下面标了精度档位与最低显存,命令行口令点右上角「复制」。

国内镜像(推荐)

civitai.work 镜像站(第三方)
页面与官网同源;下载同样需要登录

官方直连

Civitai 模型页(登录后点 Download)
浏览器已登录即可一键下载,最省事
直链 · models/3425394
需要登录态或 Token,未登录会返回 401
命令行下载(需 API Token:civitai.com/user/account)
curl -L -o "FullFineBody_MossImg_v1.safetensors" "https://civitai.com/api/download/models/3425394?token=<你的TOKEN>"
多线程断点续传(aria2)
aria2c -x8 -s8 -c -o "FullFineBody_MossImg_v1.safetensors" "https://civitai.com/api/download/models/3425394?token=<你的TOKEN>"

显存是「最低可跑」的估算:ComfyUI 会分时加载权重与文本编码器,所以最低显存≈权重体积再上一档;实际还要给激活值留余量,视频模型另受时长与分辨率影响。

放到哪:底模 → ComfyUI/models/diffusion_models;LoRA / 微调 → models/loras 或 models/checkpoints;文本编码器 → models/text_encoders;VAE → models/vae。Comfy-Org 的仓已按这套目录结构打包,--local-dir ComfyUI\models 即可直接落位。