使用Hugging Face DiffusionPipeline时遇TypeError问题求助

问题描述 #

尝试使用Hugging Face的DiffusionPipeline实现文本生成图像功能，在导入阶段触发错误：TypeError: expected string or bytes-like object，完整报错回溯如下：

The traceback:
 File "c:\Users\BHAVYA SHAH\Desktop\Artificial Intelligence\app.py", line 9, in <module>
    from diffusers import DiffusionPipeline
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\diffusers\__init__.py", line 26, in <module>
    from .pipelines import DDIMPipeline, DDPMPipeline, KarrasVePipeline, LDMPipeline, PNDMPipeline, ScoreSdeVePipeline
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\diffusers\pipelines\__init__.py", line 11, in <module>
    from .latent_diffusion import LDMTextToImagePipeline
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\diffusers\pipelines\latent_diffusion\__init__.py", line 6, in <module>
    from .pipeline_latent_diffusion import LDMBertModel, LDMTextToImagePipeline
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\diffusers\pipelines\latent_diffusion\pipeline_latent_diffusion.py", line 9, in <module>
    from transformers.activations import ACT2FN
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\transformers\__init__.py", line 30, in <module>
    from . import dependency_versions_check
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\transformers\dependency_versions_check.py", line 17, in <module>
    from .utils.versions import require_version, require_version_core
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\transformers\utils\__init__.py", line 34, in <module>
    from .generic import (
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\transformers\utils\generic.py", line 29, in <module>
    from .import_utils import is_flax_available, is_tf_available, is_torch_available, is_torch_fx_proxy
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\transformers\utils\import_utils.py", line 376, in <module>
    torch_version = version.parse(importlib_metadata.version("torch"))
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\packaging\version.py", line 49, in parse
    return Version(version)
  File "C:\Users\BHAVYA SHAH\anaconda3\lib\site-packages\packaging\version.py", line 264, in __init__ 
    match = self._regex.search(version)
TypeError: expected string or bytes-like object

解决方案 #

从报错回溯可判断，问题根源是importlib_metadata.version("torch")未返回有效字符串版本号，大概率是Torch安装损坏或依赖库版本不兼容导致，按以下步骤处理：

• 重新安装Torch
  先卸载当前可能损坏的Torch安装：

  pip uninstall torch -y
  

  根据硬件环境安装对应版本的Torch：

  • 若使用CUDA 11.8：

    pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
    

  • 若使用CPU：

    pip install torch torchvision torchaudio
    

• 验证Torch版本获取是否正常
  运行以下代码检查版本读取是否正常：

  import importlib_metadata
  print(importlib_metadata.version("torch"))
  

  若能正常输出版本号，说明Torch安装修复成功；若仍报错，继续下一步。

• 更新版本解析依赖库
  packaging和importlib_metadata负责版本解析，版本过旧可能引发异常：

  pip install --upgrade packaging importlib_metadata
  

• 更新核心依赖库
  确保Diffusers和Transformers为最新兼容版本：

  pip install --upgrade diffusers transformers
  

内容的提问来源于stack exchange，提问作者Bhavya Shah