This is an automated email from the ASF dual-hosted git repository.
reminisce pushed a change to branch subgraph
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 74479b8 [MXNET-386] ongoing maintenance on NDArray (#11126)
add fc102c3 Fix build.py when CCACHE_DIR is set. (#11226)
add c54f9f7 [MXNET-534] Add examples to example list (#11225)
add bf26886 gpu mem pool strategy (#11041)
add eb95d7b [MXNET-543] disable scalatest on Spark (#11264)
add d79e1ad [MXNET-290] MKLDNN support for model quantization (#10433)
add 66ab27e add import_ for SymbolBlock (#11127)
add bbc7a22 Scala inference memory leak fix (#11204)
add 60d7090 [MXNET-521] Add Facebook open-graph tag integration (#11155)
add a40055e [MXNET-546] Add unit test for MKLDNNSum (#11272)
add ee1a10f Improve hybridblock doc (#11236)
add 4a81433 disable testcase test_gru_bidirectional temporarily (#11273)
add 496b0f2 [MXNET-532] Clarify documentation of save_parameters(),
load_parameters() (#11210)
add 2eadfa2 Enable shared ccache and fix test (#11269)
add b819fd9 Improve data transform for gluon data loader (#11183)
add 17803b6 [ARM] improvements to ARMv7 based builds. (#11245)
add a7952f0 [MXNET-538] Add XUnit to python tests (#11199)
add feef067 [BUILD] Remove useless additional link opt (#11300)
add ebf3777 [MXNET-57] Add android64 build (#11188) (#11055)
add fb084cc leaky relu speed (#11012)
add 258e96d [MXNET-541] Correcting the new install pages bugs (#11261)
add e48a8fd Revert "[WIP] Do Not Merge. Static memory allocation for
cached_op (#10817)" (#11311)
add 02e8a71 [MXNET-539] Allow Scala users to specify data/label names for
NDArrayIter (#11256)
add 5431e12 Static alloc for hybridblock (#11313)
add 9b8eb56 Revert "Static alloc for hybridblock (#11313)" (#11318)
add 7210c2c bump up rtol for fp16 case of test_sgd (#11246)
add 11dd56a [MXNET-352] Document behavior of mx.initializer.Constant
(#10637)
add ef3169a [MXNET-554] Add ccache reporting to CI (#11322)
add 7975265 [MXNET-540] update python gpu build from source instructions
on windows (#11276)
add 16a06b6 Fix gperftools/jemalloc and lapack warning bug. (#11110)
add 0910450 allow user to define unknown token symbol (#10461)
add 92fde19 [MXNET-542] Fix mkldnn performance regression + improve test
logging (#11262)
add 0ac9f8b Bring back MXNET_GPU_COPY_NTHREADS env variable (#11237)
add 6b5d9f9 Fix the inaccruate instruction on installing nose (#11317)
add f34dafe [MXNET-380] count_include_pad argument for Avg Pooling
(#11021)
add c2714c0 MXNET_FORCE_ADDTAKEGRAD to disable
AddTakeGradLargeBatchCaller (#11316)
add 9226d91 [MXNET-514] Add clip_global_norm(row_sparse_grad). Fix
row_sparse_param.save(). Fix trainer init_kvstore (#11266)
add 00681c3 [MXNET-547] Tutorial explaining how to use the profiler
(#11274)
add 47e2b89 [MXNET-553] Restructure dockcross dockerfiles to fix caching
(#11302)
add ccee176 Add standard ResNet data augmentation for ImageRecordIter
(#11027)
add 6307c00 Fix axis Bug in MKLDNN Softmax (#11335)
add 12cb4b9 [MXNET-498] Test MKLDNN backward operators (#11232)
add a7ea976 Static alloc for hybridblock (#11320)
add 7aa8a0c Fix save load doc (#11345)
add ba9784d [MXNET-404] elemwise_add/sub between rsp and rsp on GPU
(#11179)
add 9ffc03c Add seed_aug parameter for ImageRecordItr to fix random seed
for default augmentation (#11247)
No new revisions were added by this update.
Summary of changes:
3rdparty/tvm | 2 +-
CMakeLists.txt | 26 +-
Jenkinsfile | 276 +++++++----
Makefile | 11 +-
ci/build.py | 19 +-
ci/docker/Dockerfile.build.android_arm64 | 23 +-
ci/docker/Dockerfile.build.android_armv7 | 16 +-
ci/docker/Dockerfile.build.arm64 | 31 +-
ci/docker/Dockerfile.build.armv6 | 26 +-
ci/docker/Dockerfile.build.armv7 | 30 +-
ci/docker/Dockerfile.build.jetson | 25 +-
ci/docker/Dockerfile.build.ubuntu_build_cuda | 4 +-
ci/docker/Dockerfile.build.ubuntu_cpu | 4 +-
ci/docker/Dockerfile.build.ubuntu_gpu | 4 +-
ci/docker/install/android_arm64_ndk.sh | 5 +-
.../docker/install/arm_openblas.sh | 14 +-
.../{ubuntu_ccache.sh => deb_ubuntu_ccache.sh} | 6 +-
.../__init__.py => ci/docker/install/ubuntu_arm.sh | 8 +-
ci/docker/runtime_functions.sh | 206 ++++++---
ci/test_docker_cache.py | 3 +
cpp-package/scripts/OpWrapperGenerator.py | 1 +
docs/_static/js/options.js | 18 +-
docs/_static/mxnet-theme/layout.html | 8 +-
docs/faq/env_var.md | 2 +-
docs/install/index.md | 174 +++++--
docs/tutorials/gluon/hybrid.md | 38 +-
docs/tutorials/gluon/naming.md | 6 +-
docs/tutorials/gluon/save_load_params.md | 16 +-
docs/tutorials/index.md | 3 +-
docs/tutorials/python/profiler.md | 205 +++++++++
example/README.md | 12 +-
example/gluon/dcgan.py | 8 +-
example/gluon/embedding_learning/train.py | 2 +-
example/gluon/image_classification.py | 8 +-
example/gluon/mnist.py | 2 +-
example/gluon/style_transfer/main.py | 8 +-
example/gluon/super_resolution.py | 4 +-
example/gluon/tree_lstm/main.py | 2 +-
example/gluon/word_language_model/train.py | 4 +-
example/image-classification/common/data.py | 48 +-
.../predict-cpp/CMakeLists.txt | 4 +-
example/image-classification/train_imagenet.py | 4 +-
example/quantization/imagenet_gen_qsym.py | 44 +-
example/quantization/imagenet_inference.py | 10 +-
include/mxnet/c_api.h | 4 +-
include/mxnet/ndarray.h | 2 +-
python/README.md | 4 +-
python/mxnet/contrib/quantization.py | 21 +-
python/mxnet/gluon/block.py | 156 ++++++-
python/mxnet/gluon/data/vision/transforms.py | 31 +-
python/mxnet/gluon/model_zoo/vision/alexnet.py | 2 +-
python/mxnet/gluon/model_zoo/vision/densenet.py | 2 +-
python/mxnet/gluon/model_zoo/vision/inception.py | 2 +-
python/mxnet/gluon/model_zoo/vision/mobilenet.py | 4 +-
python/mxnet/gluon/model_zoo/vision/resnet.py | 4 +-
python/mxnet/gluon/model_zoo/vision/squeezenet.py | 2 +-
python/mxnet/gluon/model_zoo/vision/vgg.py | 4 +-
python/mxnet/gluon/nn/conv_layers.py | 22 +-
python/mxnet/gluon/parameter.py | 24 +-
python/mxnet/gluon/trainer.py | 2 +-
python/mxnet/gluon/utils.py | 8 +-
python/mxnet/initializer.py | 12 +-
python/mxnet/rnn/io.py | 10 +-
python/mxnet/test_utils.py | 22 +-
.../main/scala/org/apache/mxnet/FeedForward.scala | 25 +-
.../scala/org/apache/mxnet/io/NDArrayIter.scala | 134 ++++--
.../src/test/scala/org/apache/mxnet/IOSuite.scala | 8 +-
scala-package/spark/pom.xml | 18 -
.../org/apache/mxnet/spark/MXNetGeneralSuite.scala | 36 +-
src/c_api/c_api_symbolic.cc | 5 +-
src/engine/threaded_engine_perdevice.cc | 6 +-
src/imperative/cached_op.cc | 6 +-
src/io/image_aug_default.cc | 253 ++++++++++-
src/operator/leaky_relu-inl.h | 132 ++++--
src/operator/mshadow_op.h | 2 +
src/operator/nn/convolution-inl.h | 2 +
src/operator/nn/convolution.cc | 2 +-
src/operator/nn/cudnn/cudnn_pooling-inl.h | 8 +-
src/operator/nn/mkldnn/mkldnn_act.cc | 22 +-
src/operator/nn/mkldnn/mkldnn_base-inl.h | 9 +-
src/operator/nn/mkldnn/mkldnn_base.cc | 43 +-
src/operator/nn/mkldnn/mkldnn_convolution-inl.h | 77 ++++
src/operator/nn/mkldnn/mkldnn_convolution.cc | 109 ++---
src/operator/nn/mkldnn/mkldnn_pooling-inl.h | 4 +
src/operator/nn/mkldnn/mkldnn_pooling.cc | 6 +-
src/operator/nn/mkldnn/mkldnn_softmax.cc | 5 +-
src/operator/nn/mkldnn/mkldnn_sum.cc | 31 +-
src/operator/nn/pool.cuh | 83 +++-
src/operator/nn/pool.h | 75 ++-
src/operator/nn/pooling-inl.h | 31 +-
src/operator/nn/pooling.cc | 2 +-
src/operator/nn/softmax.cc | 4 +-
src/operator/operator_tune.cc | 1 +
src/operator/quantization/dequantize.cc | 24 +
.../quantization/mkldnn/mkldnn_dequantize-inl.h | 105 +++++
.../quantization/mkldnn/mkldnn_quantize-inl.h | 112 +++++
.../quantization/mkldnn/mkldnn_quantized_conv.cc | 89 ++++
.../mkldnn/mkldnn_quantized_pooling.cc | 54 +++
.../quantization/mkldnn/mkldnn_requantize-inl.h | 158 +++++++
src/operator/quantization/quantize.cc | 24 +
src/operator/quantization/quantize_graph_pass.cc | 20 +-
src/operator/quantization/quantized_conv.cc | 27 +-
src/operator/quantization/quantized_flatten-inl.h | 23 +-
src/operator/quantization/quantized_pooling.cc | 31 +-
src/operator/quantization/requantize.cc | 25 +
src/operator/random/shuffle_op.cc | 5 +-
src/operator/tensor/elemwise_binary_op-inl.h | 18 +-
src/operator/tensor/elemwise_binary_op.h | 2 +-
src/operator/tensor/elemwise_binary_op_basic.cc | 5 +
src/operator/tensor/elemwise_binary_op_basic.cu | 134 ++++++
src/operator/tensor/indexing_op.h | 6 +-
src/storage/pooled_storage_manager.h | 181 +++++++-
src/storage/storage.cc | 16 +-
tests/cpp/operator/mkldnn.cc | 358 +++++++++------
tests/cpp/storage/storage_test.cc | 36 +-
tests/python/gpu/test_forward.py | 2 +-
tests/python/gpu/test_gluon_model_zoo_gpu.py | 2 +-
tests/python/gpu/test_kvstore_gpu.py | 4 +-
tests/python/gpu/test_operator_gpu.py | 90 ++--
.../test_quantization_mkldnn.py} | 60 ++-
tests/python/quantization/test_quantization.py | 506 ++++++++++++---------
.../train/{test_dtype.py => test_resnet_aug.py} | 43 +-
tests/python/unittest/common.py | 8 +
tests/python/unittest/test_autograd.py | 2 +-
tests/python/unittest/test_contrib_autograd.py | 2 +-
tests/python/unittest/test_exc_handling.py | 2 +-
tests/python/unittest/test_executor.py | 2 +-
tests/python/unittest/test_gluon.py | 120 +++--
tests/python/unittest/test_gluon_contrib.py | 2 +-
tests/python/unittest/test_gluon_data.py | 2 +-
tests/python/unittest/test_gluon_data_vision.py | 27 +-
tests/python/unittest/test_gluon_model_zoo.py | 2 +-
tests/python/unittest/test_gluon_trainer.py | 48 +-
tests/python/unittest/test_io.py | 167 +++++--
tests/python/unittest/test_kvstore.py | 2 +-
tests/python/unittest/test_loss.py | 2 +-
tests/python/unittest/test_module.py | 2 +-
tests/python/unittest/test_ndarray.py | 7 +-
tests/python/unittest/test_operator.py | 19 +-
tests/python/unittest/test_optimizer.py | 9 +-
tests/python/unittest/test_random.py | 2 +-
tests/python/unittest/test_recordio.py | 2 +-
tests/python/unittest/test_rnn.py | 10 +-
tests/python/unittest/test_sparse_ndarray.py | 2 +-
tests/python/unittest/test_sparse_operator.py | 37 +-
tests/tutorials/test_tutorials.py | 3 +
146 files changed, 4076 insertions(+), 1412 deletions(-)
copy example/rcnn/script/additional_deps.sh =>
ci/docker/install/arm_openblas.sh (81%)
rename ci/docker/install/{ubuntu_ccache.sh => deb_ubuntu_ccache.sh} (85%)
copy example/ssd/dataset/pycocotools/__init__.py =>
ci/docker/install/ubuntu_arm.sh (92%)
create mode 100644 docs/tutorials/python/profiler.md
create mode 100644 src/operator/nn/mkldnn/mkldnn_convolution-inl.h
create mode 100644 src/operator/quantization/mkldnn/mkldnn_dequantize-inl.h
create mode 100644 src/operator/quantization/mkldnn/mkldnn_quantize-inl.h
create mode 100644 src/operator/quantization/mkldnn/mkldnn_quantized_conv.cc
create mode 100644 src/operator/quantization/mkldnn/mkldnn_quantized_pooling.cc
create mode 100644 src/operator/quantization/mkldnn/mkldnn_requantize-inl.h
copy tests/python/{quantization_gpu/test_quantization_gpu.py =>
mkl/test_quantization_mkldnn.py} (92%)
copy tests/python/train/{test_dtype.py => test_resnet_aug.py} (84%)
