ai---sd 模型头像生成

 

https://jike.city/zho

https://github.com/ZHO-ZHO-ZHO

2种方式安装自定义 节点 custom_nodes
Method 1:
Navigate to the custom_nodes directory inside the ComfyUI root folder, typically found at /workspace/ComfyUI/custom_nodes in Runpod.
Clone this repository using git clone https://github.com/ronniebasak/ComfyUI-Tara-LLM-Integration.git .
Restart ComfyUI and reload your browser to apply changes.
Method 2: (Prerequisite: ComfyUI manager)
Open ComfyUI Manager.
Click "Install via Git URL".
Paste the link of this repository ( https://github.com/ronniebasak/ComfyUI-Tara-LLM-Integration.git ).
Restart ComfyUI and reload your browser to apply changes.

 

手动安装：custom_nodes 新节点
cd custom_nodes
git clone https://github.com/ZHO-ZHO-ZHO/ComfyUI-PuLID-ZHO.git
cd custom_nodes/ComfyUI-PuLID-ZHO
pip install -r requirements.txt
重启 ComfyUI

 

完整的安装custom_nodes 包
%cd /kaggle/working/ComfyUI/custom_nodes
!git clone https://github.com/ronniebasak/ComfyUI-Tara-LLM-Integration.git
%cd /kaggle/working/ComfyUI/custom_nodes/ComfyUI-Tara-LLM-Integration
!pip install -r requirements.txt

 

# Install the custom_node
update_manager = False
%cd /kaggle/working/ComfyUI/custom_nodes
!git clone https://github.com/ltdrdata/ComfyUI-Manager.git
%cd ComfyUI-Manager
if update_manager:
get_ipython().system('git pull')
!pip install -r requirements.txt

 

 

customer_node git上的地址，如果能安装成功， comfyui上能找到节点， 案例一般在 
在源码的 examples目录中，您可以找到一些基本的工作流。

 

 

 

卡通头像

https://github.com/cubiq/PuLID_ComfyUI
https://github.com/ToTheBeginning/PuLID

 

字节换脸插件
https://news.cafa.edu.cn/MobileNews/independenWeixinContent?contentId=225334985
https://zhuanlan.zhihu.com/p/697350603

ip-adapter_pulid_sdxl_fp16.safetensors
https://huggingface.co/huchenlei/ipadapter_pulid/resolve/main/ip-adapter_pulid_sdxl_fp16.safetensors

https://mp.weixin.qq.com/s/erRpXyv3T49MA-PzRKf2vA

https://mp.weixin.qq.com/s/aWmi8T-soJ-IiUNm0cNAlw

• ip-adapter项目主页 - https://ip-adapter.github.io/

• 模型下载地址 - https://huggingface.co/h94/IP-Adapter-FaceID

https://news.cafa.edu.cn/MobileNews/independenWeixinContent?contentId=225334985
https://github.com/ToTheBeginning/PuLID?tab=readme-ov-file
https://github.com/ZHO-ZHO-ZHO/ComfyUI-PuLID-ZHO/tree/main?tab=readme-ov-file
https://github.com/cubiq/PuLID_ComfyUI/blob/main/examples/pulid_wf.jpg

 

 

新的生成人头像的 继续 安装和验证
https://mp.weixin.qq.com/s/erRpXyv3T49MA-PzRKf2vA

 

 

 

 

 

 

 

国内类似C站的 模型集合网站
https://www.liblib.art

 

 

vage 平台在下线 生成 数字人
https://www.bilibili.com/video/BV1Kh4y1v74u/?spm_id_from=333.788.recommend_more_video.0

 

 

https://github.com/ZHO-ZHO-ZHO
这个人玩了很多模型

https://github.com/ZHO-ZHO-ZHO/ComfyUI-Workflows-ZHO
这是他的所有的 工作流.json文件 (会自动下载模型)？

https://github.com/ZHO-ZHO-ZHO/comfyui-portrait-master-zh-cn
人物肖像提示词生成模块 (肖像大师)，优化肖像生成，选择永远比填空更适合人类！

由于肖像大师的本质是提示词，因此想要通过纯提示词实现姿势的稳定生成需要大量抽卡才能实现，这是我测试抽卡了一下午精选的结果，所以建议配合 openpose 实现姿势的精确控制

Model Pose Library

 

 

t他的 ：头像，人脸，模型
https://github.com/ZHO-ZHO-ZHO/ComfyUI-PuLID-ZHO
https://github.com/ZHO-ZHO-ZHO/ComfyUI-InstantID
https://github.com/ZHO-ZHO-ZHO/ComfyUI-PhotoMaker-ZHO

基础
https://github.com/ToTheBeginning/PuLID?tab=readme-ov-file
https://github.com/cubiq/PuLID_ComfyUI

 

 

@@@模型 接入提示词模型， 方便编写提示词

https://github.com/ZHO-ZHO-ZHO/ComfyUI-Qwen-2

https://github.com/ZHO-ZHO-ZHO/ComfyUI-Gemini

 

 

https://www.liblib.art/modelinfo/d7b9bbf78364498cbf85f4526c6411c4
很真实的人物模型

https://www.liblib.art/modelinfo/68b5656a75b14d9190c50f81ceaf858a