This is an automated email from the ASF dual-hosted git repository.
haoj pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 3d38dbd Adds min cuda version assertion decorator for unit tests
(#17054)
add 7dbb4b7 fix precision problem in linalg_solve, linalg_tensorinv,
linalg_cholesky op test (#16981)
No new revisions were added by this update.
Summary of changes:
python/mxnet/ndarray/numpy/linalg.py | 57 ++++++-
python/mxnet/numpy/linalg.py | 57 ++++++-
python/mxnet/numpy_dispatch_protocol.py | 1 +
python/mxnet/symbol/numpy/linalg.py | 57 ++++++-
src/operator/numpy/linalg/np_tensorinv-inl.h | 166 +++++++++++++++++++++
src/operator/numpy/linalg/np_tensorinv.cc | 134 +++++++++++++++++
.../numpy/linalg/{np_gesvd.cu => np_tensorinv.cu} | 14 +-
.../python/unittest/test_numpy_interoperability.py | 26 ++++
tests/python/unittest/test_numpy_op.py | 166 ++++++++++++++++++---
9 files changed, 646 insertions(+), 32 deletions(-)
create mode 100644 src/operator/numpy/linalg/np_tensorinv-inl.h
create mode 100644 src/operator/numpy/linalg/np_tensorinv.cc
copy src/operator/numpy/linalg/{np_gesvd.cu => np_tensorinv.cu} (75%)
