This is an automated email from the ASF dual-hosted git repository.
taolv pushed a change to branch mkldnn-v1.0
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 5e5fe04 [mkldnn-v1.0] Initiate the transition to MKL-DNN v1.0 (#15706)
add 8e31dad [MXNET-978] Add higher order gradient support `tan`, `tanh`
(#15253)
add 90f881d Fix USE_BLAS == openblas check (#15691)
add 93708eb Fix url (404 error) (#15683)
add 0f28f5b Add -DMXNET_USE_OPENMP to Makefiles so libinfo gets updated
accordingly (#15498)
add f0b6d72 update previous flaky naive engine test (#15651)
add f2ac85a update julia install doc (#15609)
add f0c69f5 Add missing default axis value to symbol.squeeze op (#15707)
add 776d9c4 add deconv in TRT subgraph (#15666)
add 9d7fc7c Fix Scala Symbolic API some/Some typo (#15687)
add 811d516 Add MKLDNN 4c layout to fix gluoncv se_resnext101_64x4d
(#15692)
add 42a47b1 Fix _copy_to on MKLDNN backend (#15637)
add d5fcd98 [DOC] refine autograd docs (#15109)
add 773f4dc Fix quantized concat when inputs are mixed int8 and uint8
(#15693)
add b3064c5 [MKLDNN]Enhance Quantization APIs and Tutorial (#15448)
add 90954ec make TransposeShape infer shape form both sides (#15713)
add 4b07386 [MXNET-1358] Fit api tutorial (#15353)
add 8c641b8 remove mshadow submodule
add 1434b98 import mshadow source tree
add b07211f Skip Flaky Test (#15722)
add cf28b46 Add magic method `abs` to NDArray and Symbol. (#15680)
add 87425d2 fix boolean_mask for 0-size output (#15731)
add 24cce9e prevent TRT_Logger to be destroyed before TRT engine (#14898)
add 7c26810 fix dropout mask output (#15697)
add 3255d87 fix broadcast op param (#15714)
add 3112893 Dynamic Library Loading Support (#15489)
add 67acbe5 [DOC] Fix many arguments in the doc: reshape_like,
arange_like, shape_array (#15752)
add 8c2bef5 Add Gather_nd Scatter_nd to NDArray API category doc (#15689)
add 51c07e5 [Dependency Update] Dependency update doc (#15045)
add ce62873 Add quantization support for GluonCV (#15754)
add 7186123 Revert "Dynamic Library Loading Support" (#15755)
add aadef2d Fix flaky test test_global_metric (#15756)
add 45db8ea Add matrix determinant operator in linalg (#15007)
add 07eb482 fix tvm cmake (#15781)
add a2b11ae Fix PR #15489 (Dynamic Library Loading Support) (#15760)
add be49b3b Numpy-compatible Infra (#15581)
add 79d8d86 Fix the bug of `MXEnginePushAsyncND` and `MXEnginePushSyncND`
(#15751)
add b6972bb add int8 bn mkldnn implementation and test (#15664)
add b77f524 [MXNET-1206] Support NDArray indexing with None and Ellipsis
(#13143)
add a3babc4 numpy-compatible sum (#15810)
add 19916af fix license issues (#15806)
add bfd3bb8 Refactor LibraryInitializer so it's thread safe. Fixes random
sporadical concurrency crashes. (#15762)
add 0eb213d Fix backward_clip num inputs and type of clip params (#15688)
add 44a7fca [Numpy] Numpy compatible slicing (#15798)
add 614cba3 Making Features as a singleton for improved caching (#15835)
add 57927a9 Clojure package remove source images (#15828)
add c3f5eea Numpy Tensordot and Dot Operator (#15820)
add bd32de4 fixing problem with existing Singleton Caching (#15868)
add 795990b numpy linspace (#15852)
add 05f3ae1 Large Index Support for Slice (#15593)
add 67daae7 tvm infra for op attrs (#15854)
add c81535c cuDNN support cleanup (#15812)
add 39bf4e0 Port several np ops to master (#15867)
add 11ce2a2 Add large tensor support binary arithmetic (#15785)
add 24a5cf0 Allow operators with multiple outputs in get_atomic_symbol
(#15740)
add f32b58e numpy-compatible split upstream (#15841)
add b914d0a Large tensor support for random ops (#15783)
new ced2bdb Merge remote-tracking branch 'origin' into mkldnn-v1.0
The 1 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "add" were already present in the repository and have only
been added to this reference.
Summary of changes:
.gitmodules | 3 -
3rdparty/mshadow | 1 -
3rdparty/mshadow/.gitignore | 21 +
3rdparty/mshadow/.travis.yml | 43 +
3rdparty/mshadow/CHANGES.md | 12 +
3rdparty/mshadow/CMakeLists.txt | 6 +
3rdparty/mshadow/LICENSE | 13 +
3rdparty/mshadow/README.md | 37 +
3rdparty/mshadow/cmake/Cuda.cmake | 324 +++
3rdparty/mshadow/cmake/Utils.cmake | 398 ++++
3rdparty/mshadow/cmake/mshadow.cmake | 91 +
3rdparty/mshadow/cmake/mshadowUtils.cmake | 2 +
3rdparty/mshadow/doc/Doxyfile | 2358 ++++++++++++++++++++
3rdparty/mshadow/doc/README.md | 321 +++
3rdparty/mshadow/doc/mkdoc.sh | 4 +
3rdparty/mshadow/guide/.gitignore | 3 +
3rdparty/mshadow/guide/Makefile | 37 +
3rdparty/mshadow/guide/README.md | 226 ++
3rdparty/mshadow/guide/basic.cpp | 161 ++
3rdparty/mshadow/guide/basic_stream.cu | 35 +
3rdparty/mshadow/guide/config.mk | 39 +
3rdparty/mshadow/guide/defop.cpp | 49 +
3rdparty/mshadow/guide/exp-template/.gitignore | 1 +
3rdparty/mshadow/guide/exp-template/Makefile | 20 +
3rdparty/mshadow/guide/exp-template/README.md | 340 +++
3rdparty/mshadow/guide/exp-template/exp_lazy.cpp | 45 +
.../mshadow/guide/exp-template/exp_template.cpp | 72 +
.../mshadow/guide/exp-template/exp_template_op.cpp | 92 +
3rdparty/mshadow/guide/mshadow-ps/.gitignore | 4 +
3rdparty/mshadow/guide/mshadow-ps/2-levels.png | Bin 0 -> 59413 bytes
3rdparty/mshadow/guide/mshadow-ps/Makefile | 45 +
3rdparty/mshadow/guide/mshadow-ps/README.md | 227 ++
3rdparty/mshadow/guide/mshadow-ps/config.mk | 40 +
3rdparty/mshadow/guide/mshadow-ps/dbstr.h | 35 +
.../mshadow/guide/mshadow-ps/dist_async_sum-inl.h | 124 +
.../mshadow/guide/mshadow-ps/dist_async_sum.cpp | 11 +
3rdparty/mshadow/guide/mshadow-ps/local.sh | 39 +
3rdparty/mshadow/guide/mshadow-ps/local_sum-inl.h | 119 +
3rdparty/mshadow/guide/mshadow-ps/local_sum.cpp | 4 +
3rdparty/mshadow/guide/mshadow-ps/local_sum.cu | 4 +
3rdparty/mshadow/guide/neuralnet/Makefile | 38 +
3rdparty/mshadow/guide/neuralnet/README.md | 16 +
3rdparty/mshadow/guide/neuralnet/config.mk | 35 +
3rdparty/mshadow/guide/neuralnet/convnet.cu | 282 +++
3rdparty/mshadow/guide/neuralnet/nnet.cu | 202 ++
3rdparty/mshadow/guide/neuralnet/nnet_ps.cu | 312 +++
3rdparty/mshadow/guide/neuralnet/util.h | 86 +
3rdparty/mshadow/make/README.md | 18 +
3rdparty/mshadow/make/mshadow.mk | 166 ++
3rdparty/mshadow/mshadow-ps/.gitignore | 3 +
3rdparty/mshadow/mshadow-ps/README.md | 4 +
3rdparty/mshadow/mshadow-ps/mshadow_ps.h | 358 +++
3rdparty/mshadow/mshadow-ps/ps_dist-inl.h | 126 ++
3rdparty/mshadow/mshadow-ps/ps_local-inl.h | 814 +++++++
3rdparty/mshadow/mshadow-ps/ps_rabit-inl.h | 113 +
3rdparty/mshadow/mshadow-ps/thread.h | 261 +++
3rdparty/mshadow/mshadow-ps/thread_util.h | 169 ++
3rdparty/mshadow/mshadow/README.md | 8 +
3rdparty/mshadow/mshadow/base.h | 1110 +++++++++
3rdparty/mshadow/mshadow/cuda/reduce.cuh | 120 +
3rdparty/mshadow/mshadow/cuda/tensor_gpu-inl.cuh | 828 +++++++
3rdparty/mshadow/mshadow/dot_engine-inl.h | 936 ++++++++
3rdparty/mshadow/mshadow/expr_engine-inl.h | 482 ++++
3rdparty/mshadow/mshadow/expr_scalar-inl.h | 165 ++
3rdparty/mshadow/mshadow/expression.h | 416 ++++
3rdparty/mshadow/mshadow/extension.h | 41 +
3rdparty/mshadow/mshadow/extension/broadcast.h | 165 ++
.../mshadow/extension/broadcast_with_axis.h | 258 +++
3rdparty/mshadow/mshadow/extension/channel_pool.h | 108 +
.../mshadow/mshadow/extension/channel_unpool.h | 137 ++
3rdparty/mshadow/mshadow/extension/choose.h | 90 +
3rdparty/mshadow/mshadow/extension/complex.h | 525 +++++
3rdparty/mshadow/mshadow/extension/concat.h | 194 ++
3rdparty/mshadow/mshadow/extension/crop.h | 119 +
3rdparty/mshadow/mshadow/extension/fill.h | 103 +
3rdparty/mshadow/mshadow/extension/flip.h | 132 ++
3rdparty/mshadow/mshadow/extension/implicit_gemm.h | 128 ++
3rdparty/mshadow/mshadow/extension/mask.h | 97 +
3rdparty/mshadow/mshadow/extension/mirror.h | 62 +
3rdparty/mshadow/mshadow/extension/one_hot.h | 87 +
.../mshadow/mshadow/extension/pack_col2patch.h | 154 ++
3rdparty/mshadow/mshadow/extension/pad.h | 111 +
3rdparty/mshadow/mshadow/extension/range.h | 118 +
.../mshadow/mshadow/extension/reduce_with_axis.h | 136 ++
3rdparty/mshadow/mshadow/extension/reduceto1d.h | 104 +
3rdparty/mshadow/mshadow/extension/reshape.h | 87 +
3rdparty/mshadow/mshadow/extension/slice.h | 156 ++
3rdparty/mshadow/mshadow/extension/slice_ex.h | 135 ++
3rdparty/mshadow/mshadow/extension/spatial_pool.h | 152 ++
.../mshadow/mshadow/extension/spatial_unpool.h | 135 ++
.../mshadow/extension/spatial_upsampling_nearest.h | 71 +
3rdparty/mshadow/mshadow/extension/swapaxis.h | 110 +
3rdparty/mshadow/mshadow/extension/take.h | 99 +
3rdparty/mshadow/mshadow/extension/take_grad.h | 111 +
3rdparty/mshadow/mshadow/extension/transpose.h | 200 ++
.../mshadow/mshadow/extension/unpack_patch2col.h | 151 ++
3rdparty/mshadow/mshadow/half.h | 354 +++
3rdparty/mshadow/mshadow/half2.h | 143 ++
3rdparty/mshadow/mshadow/io.h | 137 ++
3rdparty/mshadow/mshadow/logging.h | 234 ++
3rdparty/mshadow/mshadow/packet-inl.h | 413 ++++
3rdparty/mshadow/mshadow/packet/plain-inl.h | 76 +
3rdparty/mshadow/mshadow/packet/sse-inl.h | 147 ++
3rdparty/mshadow/mshadow/random.h | 570 +++++
3rdparty/mshadow/mshadow/stream_gpu-inl.h | 214 ++
3rdparty/mshadow/mshadow/tensor.h | 1081 +++++++++
3rdparty/mshadow/mshadow/tensor_container.h | 208 ++
3rdparty/mshadow/mshadow/tensor_cpu-inl.h | 627 ++++++
3rdparty/mshadow/mshadow/tensor_gpu-inl.h | 245 ++
3rdparty/mshadow/scripts/travis_script.sh | 19 +
3rdparty/mshadow/test/Makefile | 35 +
3rdparty/mshadow/test/pairtest.cu | 105 +
3rdparty/mshadow/test/pool.cu | 69 +
3rdparty/mshadow/test/reshape.cu | 74 +
3rdparty/mshadow/test/test.cu | 79 +
3rdparty/mshadow/test/test.h | 67 +
3rdparty/mshadow/test/unpack.cu | 85 +
CMakeLists.txt | 20 +-
LICENSE | 355 ++-
Makefile | 10 +-
benchmark/opperf/rules/default_params.py | 6 +-
ci/jenkins/Jenkins_steps.groovy | 10 +-
cmake/BuildTVM.cmake | 2 +-
contrib/clojure-package/.gitignore | 1 +
.../infer/objectdetector/images/marcel.jpg | Bin 116161 -> 0 bytes
.../clojure-package/scripts/get_test_images.sh | 18 +-
.../test/org/apache/clojure_mxnet/image_test.clj | 6 +-
.../clojure_mxnet/infer/imageclassifier_test.clj | 5 +-
.../test/test-images/Pug-Cookie.jpg | Bin 104323 -> 0 bytes
.../clojure-package/test/test-images/kitten.jpg | Bin 110969 -> 0 bytes
contrib/clojure-package/test/test_helper.clj | 26 +
contrib/tvmop/compile.py | 4 +-
contrib/tvmop/opdef.py | 12 +-
cpp-package/scripts/OpWrapperGenerator.py | 1 +
docs/api/python/autograd/autograd.md | 75 +-
docs/api/python/gluon/contrib.md | 30 +
docs/api/python/ndarray/ndarray.md | 2 +
docs/api/python/symbol/linalg.md | 2 +
docs/api/scala/symbol.md | 2 +-
docs/faq/env_var.md | 10 +-
docs/install/index.md | 64 +-
docs/tutorials/gluon/fit_api_tutorial.md | 271 +++
docs/tutorials/index.md | 5 +-
docs/tutorials/mkldnn/mkldnn_quantization.md | 259 +++
example/README.md | 2 +-
.../__init__.py => example/lib_api/Makefile | 16 +-
example/lib_api/libtest.cc | 78 +
example/lib_api/mylib.cc | 37 +
.../contrib/__init__.py => example/lib_api/test.py | 16 +-
example/quantization/README.md | 95 +-
example/quantization/imagenet_gen_qsym_mkldnn.py | 168 +-
include/mxnet/base.h | 4 +-
include/mxnet/c_api.h | 197 +-
include/mxnet/c_predict_api.h | 4 +-
include/mxnet/lib_api.h | 50 +
include/mxnet/ndarray.h | 6 +-
include/mxnet/tuple.h | 15 +
python/mxnet/__init__.py | 6 +
python/mxnet/_ctypes/ndarray.py | 36 +-
python/mxnet/_ctypes/symbol.py | 13 +-
python/mxnet/_numpy_op_doc.py | 54 +
python/mxnet/autograd.py | 3 +
python/mxnet/base.py | 121 +-
python/mxnet/contrib/amp/lists/symbol.py | 4 +
python/mxnet/contrib/quantization.py | 449 +++-
python/mxnet/cython/ndarray.pyx | 27 +-
python/mxnet/cython/symbol.pyx | 16 +-
python/mxnet/gluon/block.py | 49 +-
python/mxnet/gluon/contrib/__init__.py | 2 +
python/mxnet/gluon/contrib/estimator/__init__.py | 2 +
python/mxnet/gluon/contrib/estimator/estimator.py | 64 +-
.../mxnet/gluon/contrib/estimator/event_handler.py | 23 +-
python/mxnet/gluon/data/dataloader.py | 2 +
python/mxnet/gluon/parameter.py | 52 +-
python/mxnet/gluon/rnn/rnn_layer.py | 2 -
python/mxnet/gluon/utils.py | 45 +-
python/mxnet/image/image.py | 2 +-
python/mxnet/library.py | 49 +
python/mxnet/ndarray/__init__.py | 6 +-
python/mxnet/ndarray/_internal.py | 11 +-
python/mxnet/ndarray/ndarray.py | 1218 ++++++----
.../{gluon/contrib => ndarray/numpy}/__init__.py | 15 +-
.../__init__.py => ndarray/numpy/_internal.py} | 7 +-
python/mxnet/ndarray/numpy/_op.py | 684 ++++++
.../__init__.py => ndarray/numpy/_register.py} | 13 +-
.../__init__.py => ndarray/numpy/linalg.py} | 9 +-
.../__init__.py => ndarray/numpy/random.py} | 8 +-
.../numpy_extension}/__init__.py | 14 +-
.../__init__.py => ndarray/numpy_extension/_op.py} | 12 +-
.../numpy_extension/_register.py} | 12 +-
.../numpy_extension/image.py} | 7 +-
python/mxnet/ndarray/register.py | 116 +-
python/mxnet/ndarray/utils.py | 7 +
python/mxnet/{gluon/contrib => numpy}/__init__.py | 21 +-
.../contrib/estimator/__init__.py => numpy/_op.py} | 7 +-
.../contrib/__init__.py => numpy/_register.py} | 12 +-
python/mxnet/numpy/arrayprint.py | 62 +
python/mxnet/numpy/function_base.py | 115 +
.../test_runtime.py => python/mxnet/numpy/io.py | 45 +-
.../estimator/__init__.py => numpy/linalg.py} | 9 +-
python/mxnet/numpy/multiarray.py | 1855 +++++++++++++++
.../estimator/__init__.py => numpy/random.py} | 9 +-
python/mxnet/numpy/stride_tricks.py | 56 +
.../{gluon/contrib/__init__.py => numpy/utils.py} | 20 +-
.../{gluon/contrib => numpy_extension}/__init__.py | 21 +-
.../__init__.py => numpy_extension/_op.py} | 7 +-
.../__init__.py => numpy_extension/_register.py} | 12 +-
.../__init__.py => numpy_extension/image.py} | 11 +-
python/mxnet/numpy_extension/utils.py | 122 +
python/mxnet/optimizer/optimizer.py | 34 +-
python/mxnet/runtime.py | 8 +-
python/mxnet/symbol/__init__.py | 7 +-
python/mxnet/symbol/_internal.py | 10 +-
.../{gluon/contrib => symbol/numpy}/__init__.py | 17 +-
.../__init__.py => symbol/numpy/_internal.py} | 7 +-
.../estimator/__init__.py => symbol/numpy/_op.py} | 7 +-
.../__init__.py => symbol/numpy/_register.py} | 13 +-
python/mxnet/symbol/numpy/_symbol.py | 1315 +++++++++++
.../__init__.py => symbol/numpy/linalg.py} | 9 +-
.../__init__.py => symbol/numpy/random.py} | 9 +-
.../contrib => symbol/numpy_extension}/__init__.py | 14 +-
.../__init__.py => symbol/numpy_extension/_op.py} | 12 +-
.../numpy_extension/_register.py} | 13 +-
.../numpy_extension/image.py} | 7 +-
python/mxnet/symbol/register.py | 84 +-
python/mxnet/symbol/symbol.py | 119 +-
python/mxnet/test_utils.py | 62 +-
python/mxnet/util.py | 395 +++-
src/c_api/c_api.cc | 175 +-
src/c_api/c_api_common.h | 15 +-
src/c_api/c_api_executor.cc | 12 +-
src/c_api/c_api_ndarray.cc | 10 +-
src/c_api/c_api_profile.cc | 2 +-
src/c_api/c_api_symbolic.cc | 228 +-
src/common/cuda_utils.h | 16 +-
src/common/utils.h | 12 +
src/engine/threaded_engine_perdevice.cc | 4 +-
src/executor/attach_op_resource_pass.cc | 4 +-
src/imperative/imperative.cc | 8 +-
src/imperative/imperative_utils.h | 41 +-
src/initialize.cc | 241 +-
src/initialize.h | 126 ++
src/io/image_io.cc | 3 +
src/ndarray/ndarray.cc | 24 +-
src/operator/bilinear_sampler.cu | 8 +-
src/operator/contrib/boolean_mask.cc | 1 +
src/operator/contrib/boolean_mask.cu | 15 +-
src/operator/contrib/multibox_detection.cc | 4 +
src/operator/contrib/multibox_prior.cc | 3 +
src/operator/contrib/multibox_target.cc | 4 +
src/operator/contrib/tvmop/ufunc.cc | 4 +-
src/operator/cudnn_bilinear_sampler-inl.h | 9 +-
src/operator/cudnn_spatial_transformer-inl.h | 9 +-
src/operator/image/crop.cc | 1 +
src/operator/image/image_random.cc | 13 +
src/operator/image/resize.cc | 1 +
src/operator/leaky_relu.cc | 1 +
src/operator/linalg.h | 50 +-
src/operator/linalg_impl.h | 243 +-
src/operator/nn/activation.cc | 1 +
src/operator/nn/batch_norm-inl.h | 52 +-
src/operator/nn/batch_norm.cc | 1 +
src/operator/nn/batch_norm.cu | 8 +-
src/operator/nn/concat.cc | 13 +-
src/operator/nn/convolution.cc | 1 +
src/operator/nn/convolution.cu | 99 +-
src/operator/nn/cudnn/cudnn_activation-inl.h | 37 +-
src/operator/nn/cudnn/cudnn_algoreg-inl.h | 4 +-
src/operator/nn/cudnn/cudnn_batch_norm-inl.h | 46 +-
src/operator/nn/cudnn/cudnn_batch_norm.cc | 4 +-
src/operator/nn/cudnn/cudnn_batch_norm.cu | 90 -
src/operator/nn/cudnn/cudnn_convolution-inl.h | 354 +--
src/operator/nn/cudnn/cudnn_deconvolution-inl.h | 246 +-
src/operator/nn/cudnn/cudnn_pooling-inl.h | 29 +-
src/operator/nn/deconvolution.cc | 1 +
src/operator/nn/dropout-inl.h | 17 +-
src/operator/nn/dropout.cc | 6 +-
src/operator/nn/fully_connected.cc | 1 +
src/operator/nn/layer_norm.cc | 1 +
src/operator/nn/mkldnn/mkldnn_base.cc | 4 +
src/operator/nn/mkldnn/mkldnn_batch_norm-inl.h | 29 +-
src/operator/nn/mkldnn/mkldnn_transpose.cc | 3 +-
src/operator/nn/pooling.cc | 3 +-
src/operator/nn/softmax.cc | 2 +
src/operator/numpy/np_broadcast_reduce_op.h | 278 +++
src/operator/numpy/np_broadcast_reduce_op_value.cc | 154 ++
src/operator/numpy/np_broadcast_reduce_op_value.cu | 48 +
src/operator/numpy/np_dot-inl.h | 110 +
src/operator/numpy/np_dot.cc | 150 ++
src/operator/numpy/np_dot.cu | 37 +
src/operator/numpy/np_elemwise_broadcast_op.cc | 186 ++
src/operator/numpy/np_elemwise_broadcast_op.cu | 82 +
src/operator/numpy/np_init_op.cc | 114 +
src/operator/numpy/np_init_op.cu | 47 +
src/operator/numpy/np_matrix_op-inl.h | 65 +
src/operator/numpy/np_matrix_op.cc | 252 +++
src/operator/numpy/np_matrix_op.cu | 41 +
src/operator/numpy/np_tensordot_op-inl.h | 688 ++++++
src/operator/numpy/np_tensordot_op.cc | 235 ++
src/operator/numpy/np_tensordot_op.cu | 42 +
src/operator/numpy/np_true_divide.cc | 127 ++
src/operator/numpy/np_true_divide.cu | 41 +
.../mkldnn/mkldnn_quantized_batch_norm.cc | 123 +
.../quantization/mkldnn/mkldnn_quantized_concat.cc | 12 +-
src/operator/quantization/quantize_graph_pass.cc | 42 +-
src/operator/quantization/quantized_batch_norm.cc | 137 ++
src/operator/quantization/quantized_concat.cc | 12 +-
src/operator/quantization/quantized_conv.cu | 9 +-
src/operator/quantization/quantized_pooling.cu | 9 +-
src/operator/random/sample_op.cc | 2 +
src/operator/random/shuffle_op.cc | 1 +
src/operator/rnn-inl.h | 162 +-
src/operator/rnn.cc | 3 +-
src/operator/roi_pooling.cc | 4 +
src/operator/sequence_mask.cc | 3 +
src/operator/spatial_transformer-inl.h | 2 +-
src/operator/spatial_transformer.cu | 8 +-
src/operator/special_functions-inl.h | 19 +
src/operator/subgraph/tensorrt/nnvm_to_onnx-inl.h | 19 +-
src/operator/subgraph/tensorrt/nnvm_to_onnx.cc | 46 +-
src/operator/subgraph/tensorrt/onnx_to_tensorrt.cc | 35 +-
src/operator/subgraph/tensorrt/onnx_to_tensorrt.h | 66 +-
src/operator/subgraph/tensorrt/tensorrt-inl.h | 27 +-
src/operator/subgraph/tensorrt/tensorrt.cc | 4 +-
src/operator/swapaxis-inl.h | 42 +-
src/operator/swapaxis.cc | 2 +-
src/operator/tensor/broadcast_reduce_op.h | 183 +-
src/operator/tensor/broadcast_reduce_op_index.cc | 1 +
src/operator/tensor/dot.cc | 1 +
src/operator/tensor/elemwise_binary_broadcast_op.h | 1 +
.../tensor/elemwise_binary_broadcast_op_logic.cc | 6 +
src/operator/tensor/elemwise_binary_op.h | 16 +-
.../tensor/elemwise_binary_scalar_op_extended.cc | 3 +-
.../tensor/elemwise_binary_scalar_op_logic.cc | 6 +
src/operator/tensor/elemwise_unary_op.h | 6 +-
src/operator/tensor/elemwise_unary_op_basic.cc | 9 +-
src/operator/tensor/elemwise_unary_op_trig.cc | 57 +-
src/operator/tensor/indexing_op.cc | 2 +
src/operator/tensor/init_op.cc | 4 +-
src/operator/tensor/la_op-inl.h | 136 +-
src/operator/tensor/la_op.cc | 166 +-
src/operator/tensor/la_op.cu | 12 +
src/operator/tensor/la_op.h | 180 ++
src/operator/tensor/matrix_op-inl.h | 171 +-
src/operator/tensor/matrix_op.cc | 15 +-
src/operator/tensor/ordering_op.cc | 1 +
src/operator/tensor/slice-inl.h | 6 +-
src/profiler/profiler.h | 15 +-
tests/cpp/engine/threaded_engine_test.cc | 117 +-
tests/nightly/test_large_array.py | 195 +-
.../test_large_vector.py} | 69 +-
tests/python/gpu/test_operator_gpu.py | 3 +
tests/python/mkl/test_quantization_mkldnn.py | 5 +-
tests/python/mkl/test_subgraph.py | 11 +
tests/python/quantization/test_quantization.py | 136 ++
tests/python/tensorrt/test_tensorrt_deconv.py | 63 +
tests/python/unittest/test_dgl_graph.py | 2 +-
tests/python/unittest/test_gluon.py | 24 +-
tests/python/unittest/test_gluon_estimator.py | 7 +-
tests/python/unittest/test_higher_order_grad.py | 40 +-
tests/python/unittest/test_library_loading.py | 48 +
tests/python/unittest/test_metric.py | 18 +-
tests/python/unittest/test_ndarray.py | 131 +-
tests/python/unittest/test_numpy_ndarray.py | 672 ++++++
tests/python/unittest/test_numpy_op.py | 834 +++++++
tests/python/unittest/test_operator.py | 114 +-
tests/python/unittest/test_profiler.py | 87 +-
tests/python/unittest/test_runtime.py | 9 +
tests/python/unittest/test_symbol.py | 27 +-
tests/python/unittest/test_thread_local.py | 36 +
tests/tutorials/test_tutorials.py | 6 +
tools/dependencies/README.md | 298 +++
372 files changed, 38271 insertions(+), 2805 deletions(-)
delete mode 160000 3rdparty/mshadow
create mode 100644 3rdparty/mshadow/.gitignore
create mode 100644 3rdparty/mshadow/.travis.yml
create mode 100644 3rdparty/mshadow/CHANGES.md
create mode 100644 3rdparty/mshadow/CMakeLists.txt
create mode 100644 3rdparty/mshadow/LICENSE
create mode 100644 3rdparty/mshadow/README.md
create mode 100644 3rdparty/mshadow/cmake/Cuda.cmake
create mode 100644 3rdparty/mshadow/cmake/Utils.cmake
create mode 100644 3rdparty/mshadow/cmake/mshadow.cmake
create mode 100644 3rdparty/mshadow/cmake/mshadowUtils.cmake
create mode 100644 3rdparty/mshadow/doc/Doxyfile
create mode 100644 3rdparty/mshadow/doc/README.md
create mode 100755 3rdparty/mshadow/doc/mkdoc.sh
create mode 100644 3rdparty/mshadow/guide/.gitignore
create mode 100644 3rdparty/mshadow/guide/Makefile
create mode 100644 3rdparty/mshadow/guide/README.md
create mode 100644 3rdparty/mshadow/guide/basic.cpp
create mode 100644 3rdparty/mshadow/guide/basic_stream.cu
create mode 100644 3rdparty/mshadow/guide/config.mk
create mode 100644 3rdparty/mshadow/guide/defop.cpp
create mode 100644 3rdparty/mshadow/guide/exp-template/.gitignore
create mode 100644 3rdparty/mshadow/guide/exp-template/Makefile
create mode 100644 3rdparty/mshadow/guide/exp-template/README.md
create mode 100644 3rdparty/mshadow/guide/exp-template/exp_lazy.cpp
create mode 100644 3rdparty/mshadow/guide/exp-template/exp_template.cpp
create mode 100644 3rdparty/mshadow/guide/exp-template/exp_template_op.cpp
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/.gitignore
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/2-levels.png
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/Makefile
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/README.md
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/config.mk
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/dbstr.h
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/dist_async_sum-inl.h
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/dist_async_sum.cpp
create mode 100755 3rdparty/mshadow/guide/mshadow-ps/local.sh
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/local_sum-inl.h
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/local_sum.cpp
create mode 100644 3rdparty/mshadow/guide/mshadow-ps/local_sum.cu
create mode 100644 3rdparty/mshadow/guide/neuralnet/Makefile
create mode 100644 3rdparty/mshadow/guide/neuralnet/README.md
create mode 100644 3rdparty/mshadow/guide/neuralnet/config.mk
create mode 100644 3rdparty/mshadow/guide/neuralnet/convnet.cu
create mode 100644 3rdparty/mshadow/guide/neuralnet/nnet.cu
create mode 100644 3rdparty/mshadow/guide/neuralnet/nnet_ps.cu
create mode 100644 3rdparty/mshadow/guide/neuralnet/util.h
create mode 100644 3rdparty/mshadow/make/README.md
create mode 100644 3rdparty/mshadow/make/mshadow.mk
create mode 100644 3rdparty/mshadow/mshadow-ps/.gitignore
create mode 100644 3rdparty/mshadow/mshadow-ps/README.md
create mode 100644 3rdparty/mshadow/mshadow-ps/mshadow_ps.h
create mode 100644 3rdparty/mshadow/mshadow-ps/ps_dist-inl.h
create mode 100644 3rdparty/mshadow/mshadow-ps/ps_local-inl.h
create mode 100644 3rdparty/mshadow/mshadow-ps/ps_rabit-inl.h
create mode 100644 3rdparty/mshadow/mshadow-ps/thread.h
create mode 100644 3rdparty/mshadow/mshadow-ps/thread_util.h
create mode 100644 3rdparty/mshadow/mshadow/README.md
create mode 100755 3rdparty/mshadow/mshadow/base.h
create mode 100644 3rdparty/mshadow/mshadow/cuda/reduce.cuh
create mode 100755 3rdparty/mshadow/mshadow/cuda/tensor_gpu-inl.cuh
create mode 100644 3rdparty/mshadow/mshadow/dot_engine-inl.h
create mode 100644 3rdparty/mshadow/mshadow/expr_engine-inl.h
create mode 100644 3rdparty/mshadow/mshadow/expr_scalar-inl.h
create mode 100644 3rdparty/mshadow/mshadow/expression.h
create mode 100644 3rdparty/mshadow/mshadow/extension.h
create mode 100644 3rdparty/mshadow/mshadow/extension/broadcast.h
create mode 100644 3rdparty/mshadow/mshadow/extension/broadcast_with_axis.h
create mode 100644 3rdparty/mshadow/mshadow/extension/channel_pool.h
create mode 100644 3rdparty/mshadow/mshadow/extension/channel_unpool.h
create mode 100644 3rdparty/mshadow/mshadow/extension/choose.h
create mode 100644 3rdparty/mshadow/mshadow/extension/complex.h
create mode 100644 3rdparty/mshadow/mshadow/extension/concat.h
create mode 100644 3rdparty/mshadow/mshadow/extension/crop.h
create mode 100644 3rdparty/mshadow/mshadow/extension/fill.h
create mode 100644 3rdparty/mshadow/mshadow/extension/flip.h
create mode 100644 3rdparty/mshadow/mshadow/extension/implicit_gemm.h
create mode 100644 3rdparty/mshadow/mshadow/extension/mask.h
create mode 100644 3rdparty/mshadow/mshadow/extension/mirror.h
create mode 100644 3rdparty/mshadow/mshadow/extension/one_hot.h
create mode 100644 3rdparty/mshadow/mshadow/extension/pack_col2patch.h
create mode 100644 3rdparty/mshadow/mshadow/extension/pad.h
create mode 100644 3rdparty/mshadow/mshadow/extension/range.h
create mode 100644 3rdparty/mshadow/mshadow/extension/reduce_with_axis.h
create mode 100644 3rdparty/mshadow/mshadow/extension/reduceto1d.h
create mode 100644 3rdparty/mshadow/mshadow/extension/reshape.h
create mode 100644 3rdparty/mshadow/mshadow/extension/slice.h
create mode 100644 3rdparty/mshadow/mshadow/extension/slice_ex.h
create mode 100644 3rdparty/mshadow/mshadow/extension/spatial_pool.h
create mode 100644 3rdparty/mshadow/mshadow/extension/spatial_unpool.h
create mode 100644
3rdparty/mshadow/mshadow/extension/spatial_upsampling_nearest.h
create mode 100644 3rdparty/mshadow/mshadow/extension/swapaxis.h
create mode 100644 3rdparty/mshadow/mshadow/extension/take.h
create mode 100644 3rdparty/mshadow/mshadow/extension/take_grad.h
create mode 100644 3rdparty/mshadow/mshadow/extension/transpose.h
create mode 100644 3rdparty/mshadow/mshadow/extension/unpack_patch2col.h
create mode 100644 3rdparty/mshadow/mshadow/half.h
create mode 100755 3rdparty/mshadow/mshadow/half2.h
create mode 100644 3rdparty/mshadow/mshadow/io.h
create mode 100644 3rdparty/mshadow/mshadow/logging.h
create mode 100644 3rdparty/mshadow/mshadow/packet-inl.h
create mode 100644 3rdparty/mshadow/mshadow/packet/plain-inl.h
create mode 100644 3rdparty/mshadow/mshadow/packet/sse-inl.h
create mode 100644 3rdparty/mshadow/mshadow/random.h
create mode 100644 3rdparty/mshadow/mshadow/stream_gpu-inl.h
create mode 100755 3rdparty/mshadow/mshadow/tensor.h
create mode 100644 3rdparty/mshadow/mshadow/tensor_container.h
create mode 100755 3rdparty/mshadow/mshadow/tensor_cpu-inl.h
create mode 100755 3rdparty/mshadow/mshadow/tensor_gpu-inl.h
create mode 100755 3rdparty/mshadow/scripts/travis_script.sh
create mode 100644 3rdparty/mshadow/test/Makefile
create mode 100644 3rdparty/mshadow/test/pairtest.cu
create mode 100644 3rdparty/mshadow/test/pool.cu
create mode 100644 3rdparty/mshadow/test/reshape.cu
create mode 100644 3rdparty/mshadow/test/test.cu
create mode 100644 3rdparty/mshadow/test/test.h
create mode 100644 3rdparty/mshadow/test/unpack.cu
delete mode 100644
contrib/clojure-package/examples/infer/objectdetector/images/marcel.jpg
copy python/mxnet/gluon/contrib/__init__.py =>
contrib/clojure-package/scripts/get_test_images.sh (66%)
mode change 100644 => 100755
delete mode 100644 contrib/clojure-package/test/test-images/Pug-Cookie.jpg
delete mode 100644 contrib/clojure-package/test/test-images/kitten.jpg
create mode 100644 contrib/clojure-package/test/test_helper.clj
create mode 100644 docs/tutorials/gluon/fit_api_tutorial.md
create mode 100644 docs/tutorials/mkldnn/mkldnn_quantization.md
copy python/mxnet/gluon/contrib/__init__.py => example/lib_api/Makefile (77%)
create mode 100644 example/lib_api/libtest.cc
create mode 100644 example/lib_api/mylib.cc
copy python/mxnet/gluon/contrib/__init__.py => example/lib_api/test.py (74%)
create mode 100644 include/mxnet/lib_api.h
create mode 100644 python/mxnet/_numpy_op_doc.py
create mode 100644 python/mxnet/library.py
copy python/mxnet/{gluon/contrib => ndarray/numpy}/__init__.py (78%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
ndarray/numpy/_internal.py} (87%)
create mode 100644 python/mxnet/ndarray/numpy/_op.py
copy python/mxnet/{gluon/contrib/__init__.py => ndarray/numpy/_register.py}
(64%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
ndarray/numpy/linalg.py} (86%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
ndarray/numpy/random.py} (86%)
copy python/mxnet/{gluon/contrib => ndarray/numpy_extension}/__init__.py (79%)
copy python/mxnet/{gluon/contrib/__init__.py =>
ndarray/numpy_extension/_op.py} (85%)
copy python/mxnet/{gluon/contrib/__init__.py =>
ndarray/numpy_extension/_register.py} (73%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
ndarray/numpy_extension/image.py} (87%)
copy python/mxnet/{gluon/contrib => numpy}/__init__.py (56%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py => numpy/_op.py} (87%)
copy python/mxnet/{gluon/contrib/__init__.py => numpy/_register.py} (69%)
create mode 100644 python/mxnet/numpy/arrayprint.py
create mode 100644 python/mxnet/numpy/function_base.py
copy tests/python/unittest/test_runtime.py => python/mxnet/numpy/io.py (50%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py => numpy/linalg.py}
(87%)
create mode 100644 python/mxnet/numpy/multiarray.py
copy python/mxnet/{gluon/contrib/estimator/__init__.py => numpy/random.py}
(87%)
create mode 100644 python/mxnet/numpy/stride_tricks.py
copy python/mxnet/{gluon/contrib/__init__.py => numpy/utils.py} (70%)
copy python/mxnet/{gluon/contrib => numpy_extension}/__init__.py (61%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
numpy_extension/_op.py} (87%)
copy python/mxnet/{gluon/contrib/__init__.py => numpy_extension/_register.py}
(68%)
copy python/mxnet/{gluon/contrib/__init__.py => numpy_extension/image.py} (85%)
create mode 100644 python/mxnet/numpy_extension/utils.py
copy python/mxnet/{gluon/contrib => symbol/numpy}/__init__.py (70%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
symbol/numpy/_internal.py} (87%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py => symbol/numpy/_op.py}
(87%)
copy python/mxnet/{gluon/contrib/__init__.py => symbol/numpy/_register.py}
(65%)
create mode 100644 python/mxnet/symbol/numpy/_symbol.py
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
symbol/numpy/linalg.py} (86%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
symbol/numpy/random.py} (86%)
copy python/mxnet/{gluon/contrib => symbol/numpy_extension}/__init__.py (79%)
copy python/mxnet/{gluon/contrib/__init__.py => symbol/numpy_extension/_op.py}
(85%)
copy python/mxnet/{gluon/contrib/__init__.py =>
symbol/numpy_extension/_register.py} (73%)
copy python/mxnet/{gluon/contrib/estimator/__init__.py =>
symbol/numpy_extension/image.py} (87%)
create mode 100644 src/initialize.h
delete mode 100644 src/operator/nn/cudnn/cudnn_batch_norm.cu
create mode 100644 src/operator/numpy/np_broadcast_reduce_op.h
create mode 100644 src/operator/numpy/np_broadcast_reduce_op_value.cc
create mode 100644 src/operator/numpy/np_broadcast_reduce_op_value.cu
create mode 100644 src/operator/numpy/np_dot-inl.h
create mode 100644 src/operator/numpy/np_dot.cc
create mode 100644 src/operator/numpy/np_dot.cu
create mode 100644 src/operator/numpy/np_elemwise_broadcast_op.cc
create mode 100644 src/operator/numpy/np_elemwise_broadcast_op.cu
create mode 100644 src/operator/numpy/np_init_op.cc
create mode 100644 src/operator/numpy/np_init_op.cu
create mode 100644 src/operator/numpy/np_matrix_op-inl.h
create mode 100644 src/operator/numpy/np_matrix_op.cc
create mode 100644 src/operator/numpy/np_matrix_op.cu
create mode 100644 src/operator/numpy/np_tensordot_op-inl.h
create mode 100644 src/operator/numpy/np_tensordot_op.cc
create mode 100644 src/operator/numpy/np_tensordot_op.cu
create mode 100644 src/operator/numpy/np_true_divide.cc
create mode 100644 src/operator/numpy/np_true_divide.cu
create mode 100644
src/operator/quantization/mkldnn/mkldnn_quantized_batch_norm.cc
create mode 100644 src/operator/quantization/quantized_batch_norm.cc
copy tests/{python/mkl/test_quantization_mkldnn.py =>
nightly/test_large_vector.py} (67%)
create mode 100644 tests/python/tensorrt/test_tensorrt_deconv.py
create mode 100644 tests/python/unittest/test_library_loading.py
create mode 100644 tests/python/unittest/test_numpy_ndarray.py
create mode 100644 tests/python/unittest/test_numpy_op.py
