AttributeError: 'NumpyOps' object has no attribute 'cblas'

[OJcoder17]

How to reproduce the behaviour

I am trying to use a tokenizing method wtih the spacy module however I receive an error whenever I try to run it; AttributeError: 'NumpyOps' object has no attribute 'cblas'

Your Environment

• Operating System: Windows 11
• Python Version Used: latest version of python
• spaCy Version Used: spacy-3.4.0
• Environment Information: Jupyter Notebook

[polm]

Sorry you're having trouble with that, that's a weird error.

How did you install spaCy? It seems like some of the dependencies may have been installed incorrectly. I would recommend you follow the instructions in the Quickstart in a fresh environment.

Also. please copy and paste code or errors as text, not as screenshots. Also please share the output of spacy info so that we know the exact details of your environment.

[TutubanaS]

(base) C:\Windows\system32>spacy info
2022-08-27 01:22:00.570216: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cudart64_110.dll'; dlerror: cudart64_110.dll not found
2022-08-27 01:22:00.570369: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.

============================== Info about spaCy ==============================

spaCy version 3.4.1
Location C:\Users\Tutu\anaconda3\lib\site-packages\spacy
Platform Windows-10-10.0.19044-SP0
Python version 3.9.12
Pipelines en_core_web_sm (3.4.0)

(base) C:\Windows\system32>

Even though I installed the packages again, I still suffer from this issue. Once I create the model and put a string in it like nlp("Hello World!"), I get this error

[polm]

Sorry you're having trouble with this. Not sure what's going on, but maybe you're using 32 bit Python? You can check with this command:

python -c "import sys;print(\"%x\" % sys.maxsize, sys.maxsize > 2**32)"

If it says False you are using 32bit Python, which is not supported - I recommend installing 64bit Python.

If that's not the issue, then could you share the output of conda list so we can tell what version of everything is installed?

You might also try installing things in a blank environment (so not base) - base is not empty, which causes issues sometimes.