Closed
Description
Since version 5.6.0, pypy is now able to run most of the numpy and scipy suites: https://morepypy.blogspot.com/2016/11/pypy27-v56-released-stdlib-2712-support.html
It'd be nice if scikit-learn could support pypy too. At the moment scikit-learn==0.18.1 fails to import with:
~> python
Python 2.7.13 (fa3249d55d15, Mar 19 2017, 20:21:48)
[PyPy 5.7.0 with GCC 6.2.0 20160901] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>>> import sklearn
>>>> from sklearn import datasets, linear_model
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/nail/home/drolando/pg/env/site-packages/sklearn/datasets/__init__.py", line 25, in <module>
from .twenty_newsgroups import fetch_20newsgroups
File "/nail/home/drolando/pg/env/site-packages/sklearn/datasets/twenty_newsgroups.py", line 54, in <module>
from ..feature_extraction.text import CountVectorizer
File "/nail/home/drolando/pg/env/site-packages/sklearn/feature_extraction/__init__.py", line 8, in <module>
from .hashing import FeatureHasher
File "/nail/home/drolando/pg/env/site-packages/sklearn/feature_extraction/hashing.py", line 9, in <module>
from . import _hashing
File "array.pxd", line 58, in init sklearn.feature_extraction._hashing (sklearn/feature_extraction/_hashing.c:5440)
ValueError: array.array has the wrong size, try recompiling. Expected 24, got 64