Reputation:

ImportError: numpy.core.multiarray failed to import

I'm trying to run this program

import cv2
import time

cv.NamedWindow("camera", 1)

capture = cv.CaptureFromCAM(0)

while True:
    img = cv.QueryFrame(capture)
    cv.ShowImage("camera", img)
    if cv.WaitKey(10) == 27:
        break
cv.DestroyAllWindows()

But I'm having a problem with numpy, I'm using pyschopy along with opencv. The problem I keep getting is this error report:

RuntimeError: module compiled against API version 7 but this version of numpy is 6

Traceback (most recent call last):

File "C:\Users\John\Documents\EyeTracking\Programs\GetImage.py", line 1, in
import cv2
ImportError: numpy.core.multiarray failed to import

I have numpy-1.6.1-py2.7 in the psychopy folder, I'm just confused as to what is wrong?

Upvotes: 258

Answers (30)

Jon

Reputation: 1142

The problem with the other answers here is that they want you to essentially upgrade some version of your packages (i.e. pip install --upgrade numpy will upgrade your numpy to the latest)... but what if you have a hard requirement to stay at the current version you are already at?

Some of the answers will work, but only if you separately install the numpy dependencies after the rest of your dependencies. If you are attempting to set this up from scratch (i.e with python3 -m pip install -r requirements.txt in a new virtual environment), you will run into this error until you also run python3 -m pip install numpy==<the older version of numpy you need>, after having initially installed your dependencies. Which essentially installs some dependencies twice or needlessly separates them into two commands, and is not ideal.

When this happens, the root issue is that some package in the pip cache causes the incorrect mixture of C compile libs... and that will keep happening on every new install, until you clear the pip cache...

If you run python3 -m pip install --no-cache-dir -r requirements.txt (with your numpy defined in the requirements.txt), and that works, but it does not after removing the --no-cache-dir option, then the cache dir is the problem.

Simply run python3 -m pip cache purge to clear the wheel packages, then re-run python3 -m pip install -r requirements.txt. Now the packages will install correctly on the first try every time.

Upvotes: 1

Bharanidharan K

Reputation: 699

If you are using python3, the following command fixes the problem on MacOS:

python3 -m pip install numpy -I

-I in detail:

pip install --help

-I, --ignore-installed      Ignore the installed packages, overwriting them.

Upvotes: 31

Erick Giffoni

Reputation: 43

I was struggling with this error in my Raspberry 4. I was developing a Django backend that used opencv.

What solved my problem was:

sudo pip install numpy==1.23.0

Raspberry info: Linux raspberrypi 5.15.61-v7l+

Python version: 3.9

Hope this helps someone.

Upvotes: 2

ril3y

Reputation: 952

If you would like to install a specific version:

pip install numpy==1.8

Upvotes: 28

Bendy Latortue

Reputation: 401

pip install opencv-python==3.4.2.17 numpy==1.14.5

did the job for me!

Upvotes: 15

phnghue

Reputation: 1686

In case you uninstalled and reinstalled or installed update but it still not work, especially if you're packaging with cx_freeze:

Try to change your code by import numpy before import cv2.

Upvotes: 0

Jerry101

Reputation: 13357

All these varied answers suggest that many different problems appear as ImportError: numpy.core.multiarray failed to import.

==> So look for more error information before the stack traceback

E.g. steps to reproduce one such problem (these steps worked fine until numpy 1.20 was released):

# set up a virtualenv
pyenv virtualenv 3.8.7 pip-issue-9542
pyenv local pip-issue-9542

# update pip, setuptools
pip install -U pip setuptools

# install 2 pips
pip install --no-cache-dir numpy==1.19.5 cvxpy==1.1.3

# try to load cvxpy
python -c 'import cvxpy'

Output:

RuntimeError: module compiled against API version 0xe but this version of numpy is 0xd
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/__init__.py", line 18, in <module>
    from cvxpy.atoms import *
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/atoms/__init__.py", line 20, in <module>
    from cvxpy.atoms.geo_mean import geo_mean
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/atoms/geo_mean.py", line 20, in <module>
    from cvxpy.utilities.power_tools import (fracify, decompose, approx_error, lower_bound,
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/utilities/power_tools.py", line 18, in <module>
    from cvxpy.atoms.affine.reshape import reshape
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/atoms/affine/reshape.py", line 18, in <module>
    from cvxpy.atoms.affine.hstack import hstack
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/atoms/affine/hstack.py", line 18, in <module>
    from cvxpy.atoms.affine.affine_atom import AffAtom
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/atoms/affine/affine_atom.py", line 22, in <module>
    from cvxpy.cvxcore.python import canonInterface
  File "/usr/local/var/pyenv/versions/pip-issue-9542/lib/python3.8/site-packages/cvxpy/cvxcore/python/__init__.py", line 3, in <module>
    import _cvxcore
ImportError: numpy.core.multiarray failed to import

The important clue is RuntimeError: module compiled against API version 0xe but this version of numpy is 0xd.

Searching on that can get you to pip issue 9542 and cvxpy issue 1229 which explain what went wrong.

The workarounds in this case are to (a) update to cvxpy>=1.1.10, or (b) update to numpy>=1.20, or (c) update both. You'll have to cope with library changes since the original environment is not reproducible without diving into pip's --no-build-isolation world.

Upvotes: 12

Vasyl Lyashkevych

Reputation: 2222

I used Anaconda environment and had the same issue. I tried all the aforementioned approaches and, alas, it didn't help me. Accumulated the suggestions, here the way which helped me:

Delete all NumPy folders in the virtual environment or in the system if you don't use a virtual environment, for example in my case:

~/home/anaconda3/envs//lib/python/site-packages/numpy

~/home/anaconda3/envs//lib/python/site-packages/numpy.libs

~/home/anaconda3/envs//lib/python/site-packages/numpy-.dist-info
Install new Numpy with:

pip install numpy -U

Hope, it could help in the same case

Upvotes: 7

theAccountant.py

Reputation: 141

Just fixed this issue. import c2 or import numpy was not working. Uninstalled the most current version of numpy. Tried to install numpy==1.15.2 just like specified above, did not work. Tried numpy==1.19.3 IT worked. I guess not all versions work perfectly with all versions of python and dependencies. So keep uninstalling and install one that works.

Upvotes: 0

Bickky Sahani

Reputation: 363

pip install numpy==1.19.3

this did the job for me.

Upvotes: 5

dope

Reputation: 185

OK so I found a solution that worked for me when trying to get OpenCV to work with Python 3.9 on Windows 10.

This is a known issue for Windows versions past v2004.

In short, the version of NumPY that you need is v1.19.3:

pip uninstall numpy

pip install numpy==1.19.3

Or to do this in one command, use the --force-reinstall flag for pip:

pip install --force-reinstall numpy==1.19.3

Upvotes: 10

MwarukaSon

Reputation: 377

run this codes worked for me, seems to be issue with version.

pip uninstall numpy

pip install numpy==1.19.3

Upvotes: 2

Michał Gacka

Reputation: 3073

Encountered this when trying to import Pytorch. Solved it by uninstalling or removing numpy repeatedly until no version was left on my Ubuntu and installing the newest version. In my case pip encountered Access is denied permission errors all the time maybe because of conflict with conda.

Run sudo pip uninstall numpy. If it refuses to delete the package from a <location> (make sure to make it the location of that specific package, not all packages - this is irreversible), run sudo rm -r <location>. Repeat above until the first command doesn't do anything and import numpy can't find the package within Python shell.

Install numpy using pip.

Upvotes: -1

alexfigtree

Reputation: 1546

In the case that

pip install -U numpy

doesn't work (even with sudo), you may want to make sure you're using the right version of numpy. I had the same "numpy.core.multiarray failed to import" issue, but it was because I had 1.6 installed for the version of Python I was using, even though I kept installing 1.8 and assumed it was installing in the right directory.

I found the bad numpy version by using the following command in my Mac terminal:

python -c "import numpy;print(numpy.__version__);print(numpy.__file__)";

This command gave me the version and location of numpy that I was using (turned out it was 1.6.2). I went to this location and manually replaced it with the numpy folder for 1.8, which resolved my "numpy.core.multiarray failed to import" issue. Hopefully someone finds this useful!

Note: For the command, use double underscore before and after 'version' and 'file'

Upvotes: 46

Lucian

Reputation: 561

I got this same error in a conda environment, only six+ years later. The other responses were helpful, and eventually I tracked it down to this problem:

> conda list numpy
# packages in environment at [Anaconda3]:
#
# Name                    Version                   Build  Channel
numpy                     1.14.5                   pypi_0    pypi
numpy-base                1.18.1           py36hc3f5095_1
numpydoc                  0.9.1                    pypi_0    pypi

The problem was that 'numpy' was a different version than 'numpy-base'. I solved this with:

> conda install numpy-base=1.14.5

so the two of them matched each other.

Upvotes: 1

R Nanthak

Reputation: 371

I was able to solve the problem by updating my python to 3.8. I am using Macbook Air with Catalina. The problem started for me after updating TensorFlow. After updating it, the error doesn't disappear after I uninstalled and installed numpy several times.

Finally, I downloaded the python3.8 package from python website and installed. Then I installed numpy and it worked.

Upvotes: 1

EQNX

Reputation: 161

Tilde folders

In the event pip uninstall numpy and reinstallation of Numpy does not work. Review your site-packages folder for sub-folders beginning with a tilde ~

These folders relate to pip installations that got mangled and the installation was aborted part way through. The tilde folders were only ever meant to be tmp folders but ended up becoming permanent. In my case there was a file called ~mpy which was a mangled legacy Numpy folder. This led to compatibility issues and ImportErrors.

These mangled folders can safely be deleted, for further details see this answer

Upvotes: 1

Nutcracker

Reputation: 359

I had the same issue, and here's how it's solved in my case.

I tried pip install -U numpy but it didn't upgrade numpy, but conda install worked for me

ImportError: numpy.core.multiarray failed to import
admin@MacBook-Air$ pip install -U numpy
Requirement already up-to-date: numpy in /Users/admin/anaconda/lib/python2.7/site-packages
admin@MacBook-Air$ python
Python 2.7.12 |Anaconda 2.4.0 (x86_64)| (default, Jul  2 2016, 17:43:17) 
[GCC 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
>>> import numpy
numpy.version.version
>>> numpy.version.version
'1.7.1'
>>> exit
Use exit() or Ctrl-D (i.e. EOF) to exit
>>> 
admin@MacBook-Air$ 
admin@MacBook-Air$ conda install numpy
Fetching package metadata .......
Solving package specifications: ..........

Package plan for installation in environment /Users/admin/anaconda:

The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    scikit-learn-0.18.1        |      np111py27_0         4.9 MB

The following packages will be UPDATED:

    numexpr:      2.3.0-np17py27_0  --> 2.6.1-np111py27_1 
    numpy:        1.7.1-py27_2      --> 1.11.2-py27_0     
    scikit-learn: 0.14.1-np17py27_1 --> 0.18.1-np111py27_0
    scipy:        0.13.2-np17py27_1 --> 0.18.1-np111py27_0

Proceed ([y]/n)? y

Fetching packages ...
scikit-learn-0 100% |#################################################################| Time: 0:00:16 312.60 kB/s
Extracting packages ...
[      COMPLETE      ]|####################################################################################| 100%
Unlinking packages ...
[      COMPLETE      ]|####################################################################################| 100%
Linking packages ...
[      COMPLETE      ]|####################################################################################| 100%

Upvotes: 1

Rider44

Reputation: 141

I was getting the same error and the problem was solved by updating my numpy installation from 1.7.1 to 1.12.1

pip install -U numpy

The followings were my cmd sequence when the error was occurred, slightly different from the above:

$ python

Python 2.7.12 |Anaconda 4.2.0 (x86_64)| (default, Jul  2 2016, 17:43:17) 

[GCC 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00)] on darwin

Type "help", "copyright", "credits" or "license" for more information.

Anaconda is brought to you by Continuum Analytics.

Please check out: http://continuum.io/thanks and https://anaconda.org

>>> import cv2

>>> import numpy as np

>>> from matplotlib import pyplot as plt

Upvotes: 3

bio_james

Reputation: 161

I had the same error message, after trying some of the suggested solutions without success, I found that I needed to run:

pip uninstall numpy

multiple times. Each time several different files were flagged for removal, from different versions of numpy that had accumulated on my system.

Once there was nothing left to uninstall, it was a matter of

pip install numpy

Hope this helps someone!

Upvotes: 11

laxman kishore

Reputation: 1

uninstall existing numpy and install opencv-python will resolve the issue

Upvotes: 0

TURUI

Reputation: 121

you may need upgrade pip, it works for me

pip install --upgrade pip
pip install -U numpy

Upvotes: 12

Constanza Garcia

Reputation: 366

For me it was a two part. First:

pip uninstall numpy

Then @Oli Blum answer (thank you):

conda install -c conda-forge numpy

That did the trick for me

Upvotes: 0

Sahil Parekh

Reputation: 9

This helped me

sudo pip3 install --upgrade --target=/home/<username>/.local/lib/python3.6/site-packages numpy

Upvotes: 0

fakturk

Reputation: 425

In my case installing from apt solved my problem.

You can try uninstall it from pip and install from apt (if you are using ubuntu etc.)

pip3 uninstall numpy
sudo apt-get install python3-numpy

Upvotes: 2

Saeed

Reputation: 3520

I don't really understand this error but I solved this error with below.

pip install -U opencv-python

Hope it was useful!

Upvotes: 2

shan89

Reputation: 11

I Had the same error occurring as I was using the numpy version suggested by the requirements.txt in the repo. When I tried to 'import pandas as pd' this error occurred. Then the solution was to upgrade numpy version to 1.15.2 as the version suggested in the requirements was mismatching with pandas. I uninstalled the existing numpy version with pip and reinstalled the new version.

pip install numpy==1.15.2

Hope this helps someone

Upvotes: 1

egeres

Reputation: 117

I had the same error after installing python and opencv in my D: drive (C: runs on a SSD). The problem seemed to be that my execution path was inside the numpy folder. You can check if that's the issue with this code :

import numpy
numpy.version.version

In case it retrieves any error related to the execution path, execute your original code from elsewhere. That solved the could not load numpy.core.multiarray error for me

Upvotes: 1