You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by ap...@apache.org on 2019/10/29 01:25:55 UTC
[incubator-mxnet] branch v1.6.x updated (4e03e6a -> 86ed5f5)
This is an automated email from the ASF dual-hosted git repository.
apeforest pushed a change to branch v1.6.x
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 4e03e6a Disables test_bulking_operator_gpu due to flakiness (#16611)
add c0e616f C Api for simplebind, fix comment for trigoops, add atol to assert (#16585)
add c574067 Imagenet inference to nightly fix (#16599)
add 7862738 Fix python doc build issue (#16630)
add 0712f00 Faster general take (#16615)
add 8c44af4 [Gluon] Don't serialize shared parameters twice (#16582)
add e262455 Fix index overflow bug in einsum (#16589)
add 29e467b Move some subgraph verbose to MXNET_SUBGRAPH_VERBOSE=2 (#16622)
add c130cc9 add npx reshape (#16640)
add 9f21cdd RNNOp only call cuda/cudnn if GPU ctx is requested (#16632)
add 73c6b4a fix bad encode (#16641)
add 84d61a1 [Perl] - ndarray to native array conversion fix (#16635)
add d12e674 fixing broken links in multiple files - round 3 (#16634)
add 22e5ae3 add type switch to weight tensor (#16543)
add 6ab4220 numpy doc enhancement (#16637)
add ffc5392 Disable float16 test (#16643)
add 11dff51 Fix GetMKLDNNData for delay alloc (#16618)
add 9322864 Infra for tvm op runtime dispatch (#16100)
add 86ed5f5 [NumPy][Operator] NumPy operator `may_share_memory` and `shares_memory` (#16533)
No new revisions were added by this update.
Summary of changes:
CMakeLists.txt | 2 +-
Makefile | 4 +-
benchmark/python/einsum/benchmark_einsum.py | 9 +
benchmark/python/tvmop/benchmark_tvmop.py | 57 ++
ci/docker/runtime_functions.sh | 5 +-
ci/jenkins/Jenkins_steps.groovy | 20 +-
contrib/tvmop/compile.py | 13 +
contrib/tvmop/core/__init__.py | 2 +-
contrib/tvmop/core/multiarray.py | 53 ++
contrib/tvmop/opdef.py | 42 +-
contrib/tvmop/space.py | 212 +++++++
docs/python_docs/environment.yml | 19 +-
.../python/tutorials/extend/custom_layer.md | 2 +-
.../gluon_from_experiment_to_deployment.md | 4 +-
.../packages/gluon/training/fit_api_tutorial.md | 2 +-
.../tutorials/packages/ndarray/sparse/train.md | 4 +-
.../packages/ndarray/sparse/train_gluon.md | 35 +-
.../tutorials/packages/onnx/fine_tuning_gluon.md | 2 +-
.../python/tutorials/packages/viz/index.rst | 2 +-
.../backend/mkldnn/mkldnn_quantization.md | 4 +-
.../tutorials/performance/backend/profiler.md | 2 +-
.../performance/backend/tensorrt/tensorrt.md | 4 +-
docs/static_site/src/pages/api/api.html | 2 +-
.../docs/tutorials/mxnet_cpp_inference_tutorial.md | 16 +-
docs/static_site/src/pages/api/faq/float16.md | 2 +-
docs/static_site/src/pages/api/faq/perf.md | 6 +-
.../src/pages/get_started/build_from_source.md | 2 +-
include/mxnet/c_api.h | 64 +++
julia/docs/src/api/io.md | 2 +-
julia/docs/src/tutorial/char-lstm.md | 2 +-
julia/docs/src/tutorial/mnist.md | 4 +-
perl-package/AI-MXNet/lib/AI/MXNet/NDArray.pm | 6 +-
perl-package/AI-MXNet/t/test_ndarray.t | 19 +-
python/mxnet/__init__.py | 2 +
python/mxnet/_ctypes/space.py | 96 ++++
python/mxnet/_numpy_op_doc.py | 126 +++++
python/mxnet/base.py | 9 +-
python/mxnet/gluon/block.py | 35 +-
python/mxnet/gluon/parameter.py | 3 +-
python/mxnet/libinfo.py | 31 +
python/mxnet/ndarray/numpy/_op.py | 230 +++++++-
python/mxnet/ndarray/numpy/random.py | 12 +-
python/mxnet/numpy/linalg.py | 25 +
python/mxnet/numpy/multiarray.py | 622 +++++++++++++++++++--
python/mxnet/numpy/random.py | 80 ++-
python/mxnet/numpy/stride_tricks.py | 9 +
python/mxnet/numpy/utils.py | 4 +-
python/mxnet/numpy_dispatch_protocol.py | 6 +-
python/mxnet/numpy_extension/random.py | 2 +-
python/mxnet/space.py | 1 +
python/mxnet/symbol/numpy/_symbol.py | 96 +++-
python/mxnet/symbol/numpy/random.py | 8 +-
python/mxnet/symbol/symbol.py | 110 ++--
python/mxnet/{gluon/__init__.py => tvmop.py} | 33 +-
python/mxnet/util.py | 63 ++-
src/c_api/c_api.cc | 23 +
src/c_api/c_api_executor.cc | 231 ++++++--
src/ndarray/ndarray.cc | 2 +
src/operator/contrib/allclose_op-inl.h | 4 +-
src/operator/contrib/boolean_mask.cc | 2 +-
src/operator/contrib/boolean_mask.cu | 2 +-
src/operator/contrib/tvmop/dot.cc | 153 +++++
src/operator/mxnet_op.h | 12 +
src/operator/numpy/np_broadcast_reduce_op_index.cc | 11 +
src/operator/numpy/np_broadcast_reduce_op_index.cu | 3 +
src/operator/numpy/np_einsum_op-inl.h | 172 +++---
src/operator/numpy/np_einsum_op.cc | 11 +
src/operator/numpy/np_einsum_path_op-inl.h | 114 ++--
src/operator/numpy/np_matrix_op-inl.h | 54 +-
src/operator/numpy/np_matrix_op.cc | 206 ++++++-
src/operator/numpy/np_matrix_op.cu | 3 +
src/operator/numpy/np_memory_op.cc | 62 ++
.../quantize.cu => numpy/np_memory_op.cu} | 12 +-
src/operator/numpy/np_memory_op.h | 75 +++
src/operator/numpy/np_nonzero_op.cc | 3 +-
src/operator/numpy/np_nonzero_op.cu | 2 +-
src/operator/numpy/random/np_choice_op.h | 20 +-
src/operator/rnn-inl.h | 5 +
src/operator/subgraph/build_subgraph.cc | 2 +-
src/operator/tensor/indexing_op.cc | 59 +-
src/operator/tensor/indexing_op.cu | 61 +-
src/operator/tensor/indexing_op.h | 28 +-
src/operator/tvmop/op_module.cc | 12 +-
src/operator/tvmop/op_module.h | 74 +++
tests/nightly/JenkinsfileForBinaries | 10 +-
tests/nightly/test_large_array.py | 6 +-
tests/nightly/test_large_vector.py | 49 +-
tests/python/unittest/test_gluon.py | 40 ++
tests/python/unittest/test_numpy_gluon.py | 23 +
.../python/unittest/test_numpy_interoperability.py | 77 +++
tests/python/unittest/test_numpy_op.py | 203 +++++--
91 files changed, 3476 insertions(+), 577 deletions(-)
create mode 100644 benchmark/python/tvmop/benchmark_tvmop.py
create mode 100644 contrib/tvmop/core/multiarray.py
create mode 100644 contrib/tvmop/space.py
create mode 100644 python/mxnet/_ctypes/space.py
create mode 120000 python/mxnet/space.py
copy python/mxnet/{gluon/__init__.py => tvmop.py} (58%)
create mode 100644 src/operator/contrib/tvmop/dot.cc
create mode 100644 src/operator/numpy/np_memory_op.cc
copy src/operator/{quantization/quantize.cu => numpy/np_memory_op.cu} (82%)
create mode 100644 src/operator/numpy/np_memory_op.h