当前位置：首页 > news >正文

SadTalker数字人服务器部署

news 2026/4/6 11:17:06

一、单独SadTalker部署

git clone https://github.com/OpenTalker/SadTalker.gitcd SadTalker conda create -n sadtalker python=3.8conda activate sadtalkerpip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113conda install ffmpegpip install -r requirements.txt

通过音频图片生成数字人：

python inference.py --driven_audio 11.wav  --source_image 11.jpeg --enhancer gfpgan

二、安装stable-diffusion-webui

环境

python：3.11

cuda：12.1

torch：2.2.1+cu121

资源：

链接：https://pan.baidu.com/s/1GeASwRZeX81ef738WG3Kiw
提取码：yn16

解压stable环境的python到/usr/local/miniconda3/envs/

unzip stable.zip /usr/local/miniconda3/envs

解压stable-webui文件

unzip stable-diffusion-webui.zip

启动服务：

python launch.py --listen --xformers --enable-insecure-extension-access --theme dark --gradio-queue  --port=8080

生成视频：

tmp2zux_56m11-0-100_enhanced

问题：

问题1：libGL.so.1: cannot open shared object file: No such file or directory

sudo apt-get update
sudo apt-get install -y libgl1-mesa-glx

问题2： python3.10/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclCommRegister

pip uninstall torch
pip install torch==2.1.0

问题3：ModuleNotFoundError: No module named 'torchvision.transforms.functional_tensor'

pip uninstall torchvision
pip install torchvision==0.16.0

1 OSError: Can't load tokenizer for 'openai/clip-vit-large-patch14'. If you were trying to load it from https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'openai/clip-vit-large-patch14' is the correct path to a directory containing all relevant files for a CLIPTokenizer tokenizer

在sd目录下创建openai并且：

git clone https://www.modelscope.cn/AI-ModelScope/clip-vit-large-patch14.git

最终：stable-diffusion-webui/openai/clip-vit-large-patch14

2 ImportError: libGL.so.1: cannot open shared object file: No such file or dir

需要更新apt，安装libgl1-mesa-glx

apt update
apt install libgl1-mesa-glx

3 AttributeError: module 'numpy' has no attribute 'complex'.

pip install --upgrade librosa