python报错 ValueError: numpy.dtype size changed, may indicate binary incompatibility.

报错

Traceback (most recent call last):
  File "E:\DeepIPShock\deepipshock_data\deepipshock_params_window_cupy.py", line 4, in <module>
    import pandas as pd
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\__init__.py", line 22, in <module>
    from pandas.compat import is_numpy_dev as _is_numpy_dev  # pyright: ignore # noqa:F401
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\compat\__init__.py", line 25, in <module>
    from pandas.compat.numpy import (
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\compat\numpy\__init__.py", line 4, in <module>
    from pandas.util.version import Version
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\util\__init__.py", line 2, in <module>
    from pandas.util._decorators import (  # noqa:F401
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\util\_decorators.py", line 14, in <module>
    from pandas._libs.properties import cache_readonly
  File "C:\Users\zhaod\.conda\envs\torch2\lib\site-packages\pandas\_libs\__init__.py", line 13, in <module>
    from pandas._libs.interval import Interval
  File "pandas\_libs\interval.pyx", line 1, in init pandas._libs.interval
ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

原因

1. ​​根本原因​​
   Pandas 内部依赖特定版本的 Numpy，当 Numpy 版本更新后，其二进制接口（如数据类型大小）发生变化，导致已编译的 Pandas 组件无法兼容新版 Numpy。

2. ​​触发场景​​

   • Conda 环境中混合使用 conda install 和 pip install 安装包
   • 手动升级/降级了 Numpy 但未同步更新 Pandas
   • 不同 Python 环境（如虚拟环境）间库版本冲突

解决

卸载并重装。

pip uninstall numpy scipy pandas

pip install numpy scipy pandas