How can I upgrade NumPy?
PythonOpencvNumpyHomebrewAnacondaPython Problem Overview
When I installed OpenCV using Homebrew (brew), I got this problem whenever I run this command to test python -c "import cv2":
RuntimeError: module compiled against API version 9 but this version of numpy is 6
Traceback (most recent call last):
File "<string>", line 1, in <module>
ImportError: numpy.core.multiarray failed to import
I tried to upgrade NumPy, but this is confusing:
>>> import numpy
>>> print numpy.__version__
1.6.1
When I run brew to upgrade NumPy, I got this problem:
brew install -u numpy
Warning: numpy-1.9.1 already installed
When I uninstalled it:
sudo pip install numpy
Requirement already satisfied (use --upgrade to upgrade): numpy in ./anaconda/lib/python2.7/site-packages
I have followed this question and deleted Anaconda from my mac.
pip install numpy
Requirement already satisfied (use --upgrade to upgrade): numpy in /Library/Python/2.7/site-packages
But nothing have changed. How can I link the NumPy version to OpenCV?
Python Solutions
Solution 1 - Python
When you already have an older version of NumPy, use this:
pip install numpy --upgrade
If it still doesn't work, try:
pip install numpy --upgrade --ignore-installed
Solution 2 - Python
Because we have two NumPy installations in the system. One is installed by Homebrew and the second is installed by pip. So in order to solve the problem, we need to delete one and use the default NumPy install by OpenCV.
Check the path,
import numpy
print numpy.__path__
and manually delete it using rm.
Solution 3 - Python
The error you mentioned happens when you have two versions of NumPy on your system. As you mentioned, the version of NumPy you imported is still not upgraded since you tried to upgrade it through pip (it will upgrade the version existing in '/Library/Python/2.7/site-packages' ).
However Python still loads the packages from '/System/Library/Frameworks/Python.framework/Versions/2.7/Extras/lib/python/numpy' where the pre-installed packages live.
In order to upgrade that version you have to use easy_install. The other way around this problem is using virtualenv and setting up a new environment with all the requirements you need.
Solution 4 - Python
#Update numpy
For python 2
pip install numpy --upgrade
You would also needed to upgrade your tables as well for updated version of numpy. so,
pip install tables --upgrade
For python 3
pip3 install numpy --upgrade
Similarly, the tables for python3 :-
pip3 install tables --upgrade
#note: You need to check which python version are you using. pip for python 2.7+ or pip3 for python 3+
Solution 5 - Python
FYI, when you using or importing TensorFlow, a similar error may occur, like (caused by NumPy):
RuntimeError: module compiled against API version 0xa but this version of numpy is 0x9
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python2.7/dist-packages/tensorflow/__init__.py", line 23, in <module>
from tensorflow.python import *
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/__init__.py", line 60, in <module>
raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/__init__.py", line 49, in <module>
from tensorflow.python import pywrap_tensorflow
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 28, in <module>
_pywrap_tensorflow = swig_import_helper()
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow', fp, pathname, description)
ImportError: numpy.core.multiarray failed to import
Error importing tensorflow. Unless you are using bazel,
you should not try to import tensorflow from its source directory;
please exit the tensorflow source tree, and relaunch your python interpreter
from there.
I followed Elmira's and Drew's solution, sudo easy_install numpy, and it worked!
sudo easy_install numpy
Searching for numpy
Best match: numpy 1.11.3
Removing numpy 1.8.2 from easy-install.pth file
Adding numpy 1.11.3 to easy-install.pth file
Using /usr/local/lib/python2.7/dist-packages
Processing dependencies for numpy
Finished processing dependencies for numpy
After that I could use TensorFlow without error.
Solution 6 - Python
I tried doing sudo pip uninstall numpy instead, because the rm didn't work at first.
Hopefully that helps.
Uninstalling then to install it again.
Solution 7 - Python
Because you have multiple versions of NumPy installed.
Try pip uninstall numpy and pip list | grep numpy several times, until you see no output from pip list | grep numpy.
Then pip install numpy will get you the newest version of NumPy.
Solution 8 - Python
This works for me:
pip install numpy --upgrade
Solution 9 - Python
If you don't encounter any permission errors with
pip install -U numpy
try:
pip install --user -U numpy
Solution 10 - Python
After installing pytorch, I got a similar error when I used:
import torch
Removing NumPy didn't help (I actually renamed NumPy, so I reverted back after it didn't work). The following commands worked for me:
sudo pip install numpy --upgrade
sudo easy_install numpy
Solution 11 - Python
All the same.
sudo easy_install numpy
My Traceback
Searching for numpy
Best match: numpy 1.13.0
Adding numpy 1.13.0 to easy-install.pth file
Using /Library/Python/2.7/site-packages
Processing dependencies for numpy
Solution 12 - Python
If you are stuck with a machine where you don't have root access, then it is better to deal with a custom Python installation.
The Anaconda installation worked like a charm:
After installation,
> [bash]$ /xxx/devTools/python/anaconda/bin/pip list --format=columns | > grep numpy > > numpy 1.13.3 numpydoc > 0.7.0