You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So far, I've hit the fact that np.array_api no longer exists:
"""
NumPy 1.22.0 was the first version to include support for the array API standard, via a separate numpy.array_api submodule. This module was marked as experimental (it emitted a warning on import) and removed in NumPy 2.0 because full support was included in the main namespace.
"""
So I put in this stanza in the tests:
if np.__version__ < '2.0.0': # not a robust way to check version!
with warnings.catch_warnings():
warnings.simplefilter("ignore")
import numpy.array_api as xp
else:
xp = np
but hit some other snags:
E AttributeError: module 'numpy' has no attribute '_dtypes'. Did you mean: 'dtypes'?
For now, I'm going to go back to <2.0 -- but hopefully I'll be able to clean this up later, if someone doesn't beat me to it.
(maybe numpy should be pinned to <2 for now?)
The text was updated successfully, but these errors were encountered:
So far, I've hit the fact that
np.array_api
no longer exists:"""
NumPy 1.22.0 was the first version to include support for the array API standard, via a separate numpy.array_api submodule. This module was marked as experimental (it emitted a warning on import) and removed in NumPy 2.0 because full support was included in the main namespace.
"""
So I put in this stanza in the tests:
but hit some other snags:
For now, I'm going to go back to <2.0 -- but hopefully I'll be able to clean this up later, if someone doesn't beat me to it.
(maybe numpy should be pinned to <2 for now?)
The text was updated successfully, but these errors were encountered: