diff options
| author | Alex Ford <fordas@uw.edu> | 2018-10-12 13:33:43 -0700 |
|---|---|---|
| committer | Facebook Github Bot <facebook-github-bot@users.noreply.github.com> | 2018-10-12 13:41:05 -0700 |
| commit | 7a1b668283ef80098d00f621803863c492900389 (patch) | |
| tree | a2b60e758996dc081261efb12e9e97b28bb90d13 /test/run_test.py | |
| parent | 134b5d62e8490ba3764d84d06facbf79fa4c9ec3 (diff) | |
| download | pytorch-7a1b668283ef80098d00f621803863c492900389.tar.gz pytorch-7a1b668283ef80098d00f621803863c492900389.tar.bz2 pytorch-7a1b668283ef80098d00f621803863c492900389.zip | |
Implement Tensor.__cuda_array_interface__. (#11984)
Summary:
_Implements pytorch/pytorch#11914, cc: ezyang_
Implements `__cuda_array_interface__` for non-sparse cuda tensors,
providing compatibility with numba (and other cuda projects...).
Adds `numba` installation to the `xenial-cuda9` jenkins test environments via direct installation in `.jenkins/pytorch/test.sh` and numba-oriented test suite in `test/test_numba_integration.py`.
See interface reference at:
https://numba.pydata.org/numba-doc/latest/cuda/cuda_array_interface.html
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11984
Differential Revision: D10361430
Pulled By: ezyang
fbshipit-source-id: 6e7742a7ae4e8d5f534afd794ab6f54f67808b63
Diffstat (limited to 'test/run_test.py')
| -rw-r--r-- | test/run_test.py | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/test/run_test.py b/test/run_test.py index 65e47ff42a..62d32874be 100644 --- a/test/run_test.py +++ b/test/run_test.py @@ -30,6 +30,7 @@ TESTS = [ 'multiprocessing', 'nccl', 'nn', + 'numba_integration', 'optim', 'sparse', 'thd_distributed', |