广告公司宣传册设计_如何进行营销型企业网站的优化_太原seo计费管理_seo短视频网页入口引流下载

环境安装

➜  ~ sw_vers
ProductName:		macOS
ProductVersion:		15.3.2
➜  ~ pip --version
pip 24.3.1 from /opt/homebrew/lib/python3.11/site-packages/pip (python 3.11)

安装ModelScope SDK

pip install modelscope

➜  ~ modelscope download --help
Traceback (most recent call last):File "/opt/homebrew/bin/modelscope", line 5, in <module>from modelscope.cli.cli import run_cmdFile "/opt/homebrew/lib/python3.11/site-packages/modelscope/cli/cli.py", line 12, in <module>from modelscope.cli.plugins import PluginsCMDFile "/opt/homebrew/lib/python3.11/site-packages/modelscope/cli/plugins.py", line 6, in <module>from modelscope.utils.plugins import PluginsManagerFile "/opt/homebrew/lib/python3.11/site-packages/modelscope/utils/plugins.py", line 18, in <module>import pkg_resourcesFile "/opt/homebrew/lib/python3.11/site-packages/pkg_resources.py", line 29, in <module>import symbol
ModuleNotFoundError: No module named 'symbol'

重新安装setuptools

pip install --upgrade setuptools

modelscope download --help

设置环境变量

MODELSCOPE_CACHE

1. 下载整个模型repo（到默认cache地址）

modelscope download --model 'Qwen/Qwen2-7b' 

2. 下载整个模型repo到指定目录

modelscope download --model 'Qwen/Qwen2-7b' --local_dir 'path/to/dir'

Wan-AI/Wan2.1-T2V-1.3B

mkdir /Volumes/m2/Wan-AI/Wan2.1-T2V-1.3B

screen -S wan

modelscope download --model 'Wan-AI/Wan2.1-T2V-1.3B' --local_dir '/Volumes/m2/Wan-AI/Wan2.1-T2V-1.3B'

Wan-AI/Wan2.1-I2V-14B-480P

modelscope download --model 'Wan-AI/Wan2.1-I2V-14B-480P' --local_dir '/Volumes/m2/Wan-AI/Wan2.1-I2V-14B-480P'

Comfy-Org/Wan_2.1_ComfyUI_repackaged

Wan_2.1_ComfyUI_repackaged

modelscope download --model 'Comfy-Org/Wan_2.1_ComfyUI_repackaged' --local_dir '/Volumes/m2/Wan-AI/Wan_2.1_ComfyUI_repackaged'

➜  split_files ls
clip_vision  diffusion_models  text_encoders  vae
➜  split_files du -sh *
1.2G	clip_vision
269G	diffusion_models
17G	text_encoders
243M	vae

创建 软链接

➜  models rm -r diffusion_models
➜  models ln -s /Volumes/m2/Wan-AI/Wan_2.1_ComfyUI_repackaged/split_files/diffusion_models diffusion_models

➜  models rm -r clip_vision
➜  models ln -s /Volumes/m2/Wan-AI/Wan_2.1_ComfyUI_repackaged/split_files/clip_vision clip_vision➜  models rm -r text_encoders
➜  models ln -s /Volumes/m2/Wan-AI/Wan_2.1_ComfyUI_repackaged/split_files/text_encoders text_encoders➜  models rm -r vae
➜  models ln -s /Volumes/m2/Wan-AI/Wan_2.1_ComfyUI_repackaged/split_files/vae vae

工作流

image_to_video_wan_480p_example.json
image_to_video_wan_720p_example.json
text_to_video_wan.json

参考

1. 模型的下载
2. Wan_2.1_ComfyUI_repackaged