You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by re...@apache.org on 2019/10/19 06:13:41 UTC
[incubator-mxnet] branch master updated (27b3e52 -> a75ec06)
This is an automated email from the ASF dual-hosted git repository.
reminisce pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 27b3e52 image crop gpu (#16464)
add a75ec06 [Numpy] einsum (#15911)
No new revisions were added by this update.
Summary of changes:
benchmark/python/einsum/benchmark_einsum.py | 78 ++
python/mxnet/ndarray/numpy/_op.py | 241 ++++-
python/mxnet/numpy/multiarray.py | 235 ++++-
python/mxnet/numpy_dispatch_protocol.py | 3 +-
python/mxnet/symbol/numpy/_symbol.py | 129 ++-
src/operator/mxnet_op.h | 36 +
src/operator/numpy/np_dot-inl.h | 16 +-
src/operator/numpy/np_dot.cc | 4 +-
src/operator/numpy/np_einsum_op-inl.h | 1092 ++++++++++++++++++++
src/operator/numpy/np_einsum_op.cc | 370 +++++++
.../numpy/{np_cumsum.cu => np_einsum_op.cu} | 15 +-
src/operator/numpy/np_einsum_path_op-inl.h | 964 +++++++++++++++++
src/operator/numpy/np_tensordot_op-inl.h | 120 ++-
src/operator/numpy/np_tensordot_op.cc | 8 +-
.../python/unittest/test_numpy_interoperability.py | 73 ++
tests/python/unittest/test_numpy_op.py | 131 +++
16 files changed, 3478 insertions(+), 37 deletions(-)
create mode 100644 benchmark/python/einsum/benchmark_einsum.py
create mode 100644 src/operator/numpy/np_einsum_op-inl.h
create mode 100644 src/operator/numpy/np_einsum_op.cc
copy src/operator/numpy/{np_cumsum.cu => np_einsum_op.cu} (72%)
create mode 100644 src/operator/numpy/np_einsum_path_op-inl.h