conda环境下face_alignment.LandmarksType._2D AttributeError: _2D解决方法

原创

已于 2024-01-03 15:31:32 修改 · 2.3k 阅读

标签

#conda

收录于

于 2023-12-31 09:01:41 首次发布

文章讲述了在运行视频重述模型时遇到的FaceAlignment异常，原因在于LandmarksType中没有_2D属性。给出了两种解决方案：直接替换为TWO_D常量或升级face_alignment库。同时介绍了Conda环境管理器在包管理和依赖性问题上的优势和用法。

1 问题描述

运行retalking模型时，代码抛出异常，信息如下所示：

Traceback (most recent call last):
  File "D:/ml/video-retalking/inference.py", line 345, in <module>
    main()
  File "D:/ml/video-retalking/inference.py", line 81, in main
    kp_extractor = KeypointExtractor()
  File "D:\ml\video-retalking\third_part\face3d\extract_kp_videos.py", line 16, in __init__
    self.detector = face_alignment.FaceAlignment(face_alignment.LandmarksType._2D)
  File "D:\.conda\retalking\lib\enum.py", line 384, in __getattr__
    raise AttributeError(name) from None
AttributeError: _2D

2 问题分析

从异常信息可知，在face_alignment.LandmarksType中没有_2D这个属性，查看face_alignment.LandmarksType源代码，如下所示：

class LandmarksType(IntEnum):
    """Enum class defining the type of landmarks to detect.

    ``TWO_D`` - the detected points ``(x,y)`` are detected in a 2D space and follow the visible contour of the face
    ``TWO_HALF_D`` - this points represent the projection of the 3