导入h5py触发numpy AttributeError：module 'numpy' has no attribute 'typeDict'

完整报错回溯 #

AttributeError                            Traceback (most recent call last)
/home/ack253/neuroaesthetics/ava-mlsp/predict_mlsp_wide.ipynb Cell 5 line 2
1 import pandas as pd, numpy as np, os
----> 2 from kuti import image_utils as img

File ~/.local/lib/python3.8/site-packages/kuti/__init__.py:8
5 import warnings
6 warnings.filterwarnings('ignore',category=FutureWarning)
----> 8 from . import generic, tensor_ops, image_utils, generators, model_helper, applications
10 # remove tensorflow warnings
11 import logging

File ~/.local/lib/python3.8/site-packages/kuti/generic.py:10
8 import numpy as np, pandas as pd
9 import multiprocessing as mp
---> 10 import os, scipy, h5py, time, sys
11 import urllib, shutil, subprocess
12 from munch import Munch

File ~/.local/lib/python3.8/site-packages/h5py/__init__.py:46
37     _warn(("h5py is running against HDF5 {0} when it was built against {1}, "
38            "this may cause problems").format(
39             '{0}.{1}.{2}'.format(*version.hdf5_version_tuple),
40             '{0}.{1}.{2}'.format(*version.hdf5_built_version_tuple)
...
318     return Tester
--> 320 raise AttributeError("module {!r} has no attribute "
321                      "{!r}".format(__name__, attr))

AttributeError: module 'numpy' has no attribute 'typeDict'

原因分析 #

numpy.typeDict是旧版本numpy的属性，在numpy 1.20及以上版本中已被移除，替换为numpy.dtypeDict。报错是因为当前安装的h5py版本过旧，不兼容你使用的numpy新版本。

解决办法 #

• 升级h5py到兼容版本：在终端执行命令：

  pip install --upgrade h5py
  

• 锁定numpy兼容版本（可选）：如果项目依赖旧版numpy特性，可安装适配的numpy版本（如1.19.x系列）：

  pip install numpy==1.19.5
  

• 检查虚拟环境：确认当前使用的虚拟环境正确，避免全局包与虚拟环境包版本冲突。

内容的提问来源于stack exchange，提问作者Alison