Stable Diffusion 3 Medium 模型

开源SD3,中型版本,20亿参数,Stable Diffusion 3 Medium,系统内存要求32G,显卡6G。

a female character with long, flowing hair that appears to be made of ethereal, swirling patterns resembling the Northern Lights or Aurora Borealis. The background is dominated by deep blues and purples, creating a mysterious and dramatic atmosphere. The character's face is serene, with pale skin and striking features. She wears a dark-colored outfit with subtle patterns. The overall style of the artwork is reminiscent of fantasy or supernatural genres

不同尺寸模型文件区别

sd3_medium.safetensors 4.34GB, includes the MMDiT and VAE weights but does not include any text encoders, you need to download T5XXL text encoder yourself.

sd3_medium_incl_clips.safetensors 5.97GB, includes all necessary weights except for the T5XXL text encoder. It requires minimal resources, but the model's performance will differ without the T5XXL text encoder.

sd3_medium_incl_clips_t5xxlfp8.safetensors 10.9GB, contains all necessary weights, including fp8 version of the T5XXL text encoder, offering a balance between quality and resource requirements.

sd3_medium_incl_clips_t5xxlfp16.safetensors 15.8GB, contains all necessary weights, including fp16 version of the T5XXL text encoder.

T5XXL is a new text encoder from Google, it has emerged as a powerful technique in natural language processing (NLP).

Clip及t5xxl_下载

1.1 clip_g.safetensors 1.39GB

1.2 clip_l.safetensors 246MB

1.3 t5xxl_fp16.safetensors 9.79GB

1.4 t5xxl_fp8_e4m3fn.safetensors 4.89GB

其他

  • 需要更新ComfyUI到最新,先git fetch再git merge

  • 命令行进入路径 ComfyUI\python_embeded安装缺失的spandrel模块,运行

bash 复制代码
python -m pip install spandrel

参考

1\]: https://edmondyip.substack.com/p/stability-ai-released-stable-diffusion

相关推荐
修炼室13 天前
Stable Diffusion WebUI 本地部署完整教程
stable diffusion
NetX行者14 天前
Stable Diffusion:开启AI图像生成新纪元
人工智能·stable diffusion
这是一个懒人18 天前
SD和comfyui常用模型介绍和下载
stable diffusion·comfyui·模型下载
有点小帅得平哥哥20 天前
Stable Diffusion WebUI 本地部署指南(Windows 11 + RTX 4060 Ti)
stable diffusion
CoovallyAIHub23 天前
突破异常数据瓶颈!AnomalyAny:一句话+一张图,零样本生成任意异常图像
计算机视觉·stable diffusion
写代码的小阿帆24 天前
Fractal Generative Models论文阅读笔记与代码分析
论文阅读·stable diffusion·transformer
春末的南方城市25 天前
港科大&快手提出统一上下文视频编辑 UNIC,各种视频编辑任务一网打尽,还可进行多项任务组合!
人工智能·计算机视觉·stable diffusion·aigc·transformer
多恩Stone1 个月前
【Stable Diffusion 1.5 】在 Unet 中每个 Cross Attention 块中的张量变化过程
stable diffusion
今夕节度使1 个月前
ARM架构推理Stable Diffusiond
stable diffusion
一刀到底2111 个月前
comfyui利用 SkyReels-V2直接生成长视频本地部署问题总结 2 :寻找丢失的model 和工作流中 get set 方法的应用
人工智能·python·comfyui