Windows/Colab环境Detectron2导入成功但创建Detectron2LayoutModel报错

问题 #

Windows系统已完成Detectron2的安装，在Jupyter Notebook切换至Detectron2安装目录后，可成功导入detectron2和layoutparser库，但创建Detectron2LayoutModel实例时触发ImportError，提示找不到detectron2库；Colab环境测试也出现相同问题。

相关代码 #

import detectron2
import layoutparser as lp  # 补充原代码缺失的别名定义

model = lp.Detectron2LayoutModel('lp://PubLayNet/mask_rcnn_X_101_32x8d_FPN_3x/config',
                                 extra_config=["MODEL.ROI_HEADS.SCORE_THRESH_TEST", 0.5],
                                 label_map={0: "Text", 1: "Title", 2: "List", 3:"Table", 4:"Figure"})

# 或
model = lp.models.Detectron2LayoutModel('lp://PubLayNet/mask_rcnn_X_101_32x8d_FPN_3x/config',
                                 extra_config=["MODEL.ROI_HEADS.SCORE_THRESH_TEST", 0.5],
                                 label_map={0: "Text", 1: "Title", 2: "List", 3:"Table", 4:"Figure"})

报错信息 #

ImportError Traceback (most recent call last)
Cell In[53], line 1
----> 1 model = lp.models.Detectron2LayoutModel(config_path ='lp://PubLayNet/faster_rcnn_R_50_FPN_3x/config')

File C:\anaconda\lib\site-packages\layoutparser\models\base_layoutmodel.py:87, in BaseLayoutModel.new(cls, *args, **kwargs)
85 def new(cls, *args, **kwargs):
---> 87 requires_backends(cls, cls.DEPENDENCIES)
88 return super().new(cls)

File C:\anaconda\lib\site-packages\layoutparser\file_utils.py:175, in requires_backends(obj, backends)
173 name = obj.name if hasattr(obj, "name") else obj.__class__.__name__
174 if not all(BACKENDS_MAPPING[backend][0]() for backend in backends):
---> 175 raise ImportError(
176 "".join([BACKENDS_MAPPING[backend][1].format(name) for backend in backends])
177 )

ImportError:
Detectron2LayoutModel requires the detectron2 library but it was not found in your environment. Checkout the instructions on the
installation page: https://github.com/facebookresearch/detectron2/blob/master/INSTALL.md and follow the ones
that match your environment. Typically the following would work for MacOS or Linux CPU machines:
pip install 'git+https://github.com/facebookresearch/detectron2.git@v0.4#egg=detectron2'

解决方案 #

1. 对齐Jupyter内核与Detectron2安装环境 #

Jupyter默认内核可能与Detectron2所在环境不匹配，导致依赖识别失败：

• 在Jupyter单元格执行以下代码，查看当前Python解释器路径：

  import sys
  print(sys.executable)
  

• 对比安装Detectron2时的Python路径，若不一致：
  • 切换到Detectron2所在环境，执行pip install ipykernel
  • 运行python -m ipykernel install --user --name=你的环境名添加对应内核
  • 在Jupyter中切换到该内核重新运行代码

2. 修复layoutparser的依赖检测逻辑 #

layoutparser的检测脚本对Windows环境的Detectron2兼容不佳，可手动修改：

• 找到layoutparser安装目录下的file_utils.py（示例路径：C:\anaconda\lib\site-packages\layoutparser\file_utils.py）
• 定位BACKENDS_MAPPING字典中的"detectron2"条目，替换检测函数为简单的模块存在性检查：

  "detectron2": (
      lambda: importlib.util.find_spec("detectron2") is not None,
      "Detectron2LayoutModel requires the detectron2 library but it was not found in your environment...",
  ),
  

• 保存文件后重启Jupyter，重新执行代码

3. Colab环境适配 #

Colab默认Detectron2版本与layoutparser要求不匹配，需重新安装：

• 在Colab中执行以下命令：

  !pip uninstall -y detectron2
  !pip install layoutparser torch torchvision
  !pip install 'git+https://github.com/facebookresearch/detectron2.git@v0.4#egg=detectron2'
  

• 重启Colab运行时后再执行模型创建代码

4. 补全代码缺失定义 #

确保导入layoutparser时添加别名lp，否则代码会因未定义lp报错：

import layoutparser as lp

内容的提问来源于stack exchange，提问作者Anya Pilipentseva