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
Then ideally we're testing libraries covered by array-api-compat
NumPy
PyTorch
Dask
CuPy is probably a no-go because GPUs on CI is too finnicky
I think it'd be really nice to test jax.experimental.array_api too as it has some distinct properties that we want to be mindful with for future test suite development.
The text was updated successfully, but these errors were encountered:
For array-api-compat I have a script that I can run on a CUDA machine that runs the CuPy tests. We could have something like that here. We could also do that for PyTorch GPU. We would then just have to run it ourselves manually from time to time.
numpy.array_api
witharray-api-strict
numpy.array_api
is going to be removed and supplanted by this third-party librarymeta/
tests (we never did before!)ARRAY_API_TESTS_VERSION
values tooarray-api-compat
jax.experimental.array_api
too as it has some distinct properties that we want to be mindful with for future test suite development.The text was updated successfully, but these errors were encountered: