所有的依赖跟模型都下载好了以后,运行 python -m scripts.inference --inference_config configs/inference/test.yaml直接报错 请问是缺少什么依赖了吗?

[zhaorongxi]

(base) (venv) PS D:\museTalk\MuseTalk> python -m scripts.inference --inference_config configs/inference/test.yaml
add ffmpeg to path
Traceback (most recent call last):
File "D:\soft\python3.10\lib\runpy.py", line 196, in run_module_as_main
return run_code(code, main_globals, None,
File "D:\soft\python3.10\lib\runpy.py", line 86, in run_code
exec(code, run_globals)
File "D:\museTalk\MuseTalk\scripts\inference.py", line 12, in
from musetalk.utils.utils import get_file_type,get_video_fps,datagen
File "D:\museTalk\MuseTalk\musetalk\utils\utils.py", line 16, in
from musetalk.models.vae import VAE
File "D:\museTalk\MuseTalk\musetalk\models\vae.py", line 1, in
from diffusers import AutoencoderKL
File "D:\museTalk\venv\lib\site-packages\diffusers_init.py", line 5, in
from .utils import (
File "D:\museTalk\venv\lib\site-packages\diffusers\utils_init.py", line 38, in
from .dynamic_modules_utils import get_class_from_dynamic_module
File "D:\museTalk\venv\lib\site-packages\diffusers\utils\dynamic_modules_utils.py", line 28, in
from huggingface_hub import cached_download, hf_hub_download, model_info
ImportError: cannot import name 'cached_download' from 'huggingface_hub' (D:\museTalk\venv\lib\site-packages\huggingface_hub_init.py)

[Amoukai]

我也有遇到，以下是詢問Claude 3.7 的回答，我執行之後成功

解決方案
降級 huggingface_hub 到包含 cached_download 函數的版本：

bash
pip uninstall huggingface_hub -y
pip install huggingface_hub==0.13.4
這個版本通常與 diffusers 0.18.0 兼容良好。如果您的 diffusers 版本不同，可能也需要調整：

bash
pip uninstall diffusers -y
pip install diffusers==0.18.0
確保相關依賴的兼容性
為了處理可能的依賴樹問題，建議還檢查以下包的版本：

bash
pip install transformers==4.29.0
pip install accelerate==0.20.3