ValueError while Importing Word2Vec from gensim.models

[usamashami11]

Description of Issue

A week ago everything was perfectly working, however, now, while importing gensim model Word2Vec, after successful installation of the gensim package in Google Colab, I am facing this error. ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

Steps to Reproduce Error

1. Install the library via !pip install gensim
2. Import Word2Vec via from gensim.models import Word2Vec

Complete Error Output

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
[<ipython-input-3-d83094a7b397>](https://localhost:8080/#) in <cell line: 0>()
      1 # Import Libraries
----> 2 from gensim.models import Word2Vec
      3 import nltk
      4 nltk.download('punkt_tab')
      5 from nltk.tokenize import word_tokenize

5 frames
[/usr/local/lib/python3.11/dist-packages/gensim/__init__.py](https://localhost:8080/#) in <module>
      9 import logging
     10 
---> 11 from gensim import parsing, corpora, matutils, interfaces, models, similarities, utils  # noqa:F401
     12 
     13 

[/usr/local/lib/python3.11/dist-packages/gensim/corpora/__init__.py](https://localhost:8080/#) in <module>
      4 
      5 # bring corpus classes directly into package namespace, to save some typing
----> 6 from .indexedcorpus import IndexedCorpus  # noqa:F401 must appear before the other classes
      7 
      8 from .mmcorpus import MmCorpus  # noqa:F401

[/usr/local/lib/python3.11/dist-packages/gensim/corpora/indexedcorpus.py](https://localhost:8080/#) in <module>
     12 import numpy
     13 
---> 14 from gensim import interfaces, utils
     15 
     16 logger = logging.getLogger(__name__)

[/usr/local/lib/python3.11/dist-packages/gensim/interfaces.py](https://localhost:8080/#) in <module>
     17 import logging
     18 
---> 19 from gensim import utils, matutils
     20 
     21 

[/usr/local/lib/python3.11/dist-packages/gensim/matutils.py](https://localhost:8080/#) in <module>
   1032 try:
   1033     # try to load fast, cythonized code if possible
-> 1034     from gensim._matutils import logsumexp, mean_absolute_difference, dirichlet_expectation
   1035 
   1036 except ImportError:

/usr/local/lib/python3.11/dist-packages/gensim/_matutils.pyx in init gensim._matutils()

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

Output of Versions

Linux-6.1.85+-x86_64-with-glibc2.35
Python 3.11.11 (main, Dec  4 2024, 08:55:07) [GCC 11.4.0]
Bits 64
NumPy 2.0.2
SciPy 1.14.1
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
[<ipython-input-5-824570de713c>](https://localhost:8080/#) in <cell line: 0>()
      4 import numpy; print("NumPy", numpy.__version__)
      5 import scipy; print("SciPy", scipy.__version__)
----> 6 import gensim; print("gensim", gensim.__version__)
      7 from gensim.models import word2vec;print("FAST_VERSION", word2vec.FAST_VERSION)

5 frames
[/usr/local/lib/python3.11/dist-packages/gensim/__init__.py](https://localhost:8080/#) in <module>
      9 import logging
     10 
---> 11 from gensim import parsing, corpora, matutils, interfaces, models, similarities, utils  # noqa:F401
     12 
     13 

[/usr/local/lib/python3.11/dist-packages/gensim/corpora/__init__.py](https://localhost:8080/#) in <module>
      4 
      5 # bring corpus classes directly into package namespace, to save some typing
----> 6 from .indexedcorpus import IndexedCorpus  # noqa:F401 must appear before the other classes
      7 
      8 from .mmcorpus import MmCorpus  # noqa:F401

[/usr/local/lib/python3.11/dist-packages/gensim/corpora/indexedcorpus.py](https://localhost:8080/#) in <module>
     12 import numpy
     13 
---> 14 from gensim import interfaces, utils
     15 
     16 logger = logging.getLogger(__name__)

[/usr/local/lib/python3.11/dist-packages/gensim/interfaces.py](https://localhost:8080/#) in <module>
     17 import logging
     18 
---> 19 from gensim import utils, matutils
     20 
     21 

[/usr/local/lib/python3.11/dist-packages/gensim/matutils.py](https://localhost:8080/#) in <module>
   1032 try:
   1033     # try to load fast, cythonized code if possible
-> 1034     from gensim._matutils import logsumexp, mean_absolute_difference, dirichlet_expectation
   1035 
   1036 except ImportError:

/usr/local/lib/python3.11/dist-packages/gensim/_matutils.pyx in init gensim._matutils()

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject

Alternative Method (1)

1. Install the library via !pip install --upgrade gensim

Output of Versions for this method

Linux-6.1.85+-x86_64-with-glibc2.35
Python 3.11.11 (main, Dec  4 2024, 08:55:07) [GCC 11.4.0]
Bits 64
NumPy 2.0.2
SciPy 1.13.1
---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
[<ipython-input-2-79ea32cbff8d>](https://localhost:8080/#) in <cell line: 0>()
      4 import numpy; print("NumPy", numpy.__version__)
      5 import scipy; print("SciPy", scipy.__version__)
----> 6 import gensim; print("gensim", gensim.__version__)
      7 from gensim.models import word2vec; print("FAST_VERSION", word2vec.FAST_VERSION)

9 frames
[/usr/local/lib/python3.11/dist-packages/gensim/__init__.py](https://localhost:8080/#) in <module>
      9 import logging
     10 
---> 11 from gensim import parsing, corpora, matutils, interfaces, models, similarities, utils  # noqa:F401
     12 
     13 

[/usr/local/lib/python3.11/dist-packages/gensim/parsing/__init__.py](https://localhost:8080/#) in <module>
      2 
      3 from .porter import PorterStemmer  # noqa:F401
----> 4 from .preprocessing import (  # noqa:F401
      5     preprocess_documents,
      6     preprocess_string,

[/usr/local/lib/python3.11/dist-packages/gensim/parsing/preprocessing.py](https://localhost:8080/#) in <module>
     24 import glob
     25 
---> 26 from gensim import utils
     27 from gensim.parsing.porter import PorterStemmer
     28 

[/usr/local/lib/python3.11/dist-packages/gensim/utils.py](https://localhost:8080/#) in <module>
     33 
     34 import numpy as np
---> 35 import scipy.sparse
     36 from smart_open import open
     37 

[/usr/local/lib/python3.11/dist-packages/scipy/sparse/__init__.py](https://localhost:8080/#) in <module>
    292 import warnings as _warnings
    293 
--> 294 from ._base import *
    295 from ._csr import *
    296 from ._csc import *

[/usr/local/lib/python3.11/dist-packages/scipy/sparse/_base.py](https://localhost:8080/#) in <module>
      3 
      4 import numpy as np
----> 5 from scipy._lib._util import VisibleDeprecationWarning
      6 
      7 from ._sputils import (asmatrix, check_reshape_kwargs, check_shape,

[/usr/local/lib/python3.11/dist-packages/scipy/_lib/_util.py](https://localhost:8080/#) in <module>
     16 
     17 import numpy as np
---> 18 from scipy._lib._array_api import array_namespace
     19 
     20 

[/usr/local/lib/python3.11/dist-packages/scipy/_lib/_array_api.py](https://localhost:8080/#) in <module>
     15 
     16 from scipy._lib import array_api_compat
---> 17 from scipy._lib.array_api_compat import (
     18     is_array_api_obj,
     19     size,

[/usr/local/lib/python3.11/dist-packages/scipy/_lib/array_api_compat/numpy/__init__.py](https://localhost:8080/#) in <module>
----> 1 from numpy import *
      2 
      3 # from numpy import * doesn't overwrite these builtin names
      4 from numpy import abs, max, min, round
      5 

[/usr/local/lib/python3.11/dist-packages/numpy/__init__.py](https://localhost:8080/#) in __getattr__(attr)
    362         try:
    363             x = ones(2, dtype=float32)
--> 364             if not abs(x.dot(x) - float32(2.0)) < 1e-5:
    365                 raise AssertionError()
    366         except AssertionError:

ModuleNotFoundError: No module named 'numpy.rec'

Alternative Method (2)

1. Install gensim from the source code (available in releases) via the following:

# Downloading tar.gz file
!wget https://github.com/piskvorky/gensim/archive/refs/tags/4.3.2.tar.gz

# Unzipping tar.gz file
!tar -xvzf /content/4.3.2.tar.gz

# Changing directory and Installing the Gensim package
import os 
os.chdir('/content/gensim-4.3.2/')
!pwd
!pip install .

Output of Versions for this method

Linux-6.1.85+-x86_64-with-glibc2.35
Python 3.11.11 (main, Dec  4 2024, 08:55:07) [GCC 11.4.0]
Bits 64
NumPy 2.0.2
SciPy 1.14.1
---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
[<ipython-input-4-824570de713c>](https://localhost:8080/#) in <cell line: 0>()
      4 import numpy; print("NumPy", numpy.__version__)
      5 import scipy; print("SciPy", scipy.__version__)
----> 6 import gensim; print("gensim", gensim.__version__)
      7 from gensim.models import word2vec;print("FAST_VERSION", word2vec.FAST_VERSION)

4 frames
[/content/gensim-4.3.2/gensim/__init__.py](https://localhost:8080/#) in <module>
      9 import logging
     10 
---> 11 from gensim import parsing, corpora, matutils, interfaces, models, similarities, utils  # noqa:F401
     12 
     13 

[/content/gensim-4.3.2/gensim/corpora/__init__.py](https://localhost:8080/#) in <module>
      4 
      5 # bring corpus classes directly into package namespace, to save some typing
----> 6 from .indexedcorpus import IndexedCorpus  # noqa:F401 must appear before the other classes
      7 
      8 from .mmcorpus import MmCorpus  # noqa:F401

[/content/gensim-4.3.2/gensim/corpora/indexedcorpus.py](https://localhost:8080/#) in <module>
     12 import numpy
     13 
---> 14 from gensim import interfaces, utils
     15 
     16 logger = logging.getLogger(__name__)

[/content/gensim-4.3.2/gensim/interfaces.py](https://localhost:8080/#) in <module>
     17 import logging
     18 
---> 19 from gensim import utils, matutils
     20 
     21 

[/content/gensim-4.3.2/gensim/matutils.py](https://localhost:8080/#) in <module>
     18 import scipy.sparse
     19 from scipy.stats import entropy
---> 20 from scipy.linalg import get_blas_funcs, triu
     21 from scipy.linalg.lapack import get_lapack_funcs
     22 from scipy.special import psi  # gamma function utils

ImportError: cannot import name 'triu' from 'scipy.linalg' (/usr/local/lib/python3.11/dist-packages/scipy/linalg/__init__.py)

I have also tried reinstalling different versions of numpy, scipy and/or gensim, but no luck. Please look into this and resolve asap. Thanks!

[gojomo]

Per my comment at #3605 (comment), I don't see a problem if (a) installing gensim via !pip install --upgrade gensim; then (b) restarting the session as is often necessary when the underlying installed packages have changed.

As this seems to be roughly the same issues as #3605, let's do further discussion there unless the errors are shown to require different workarounds - so closing this as duplicate for now.