You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by la...@apache.org on 2020/11/05 02:04:30 UTC
[incubator-mxnet] branch leezu-patch-2 updated (9fedcf4 -> bae10e6)
This is an automated email from the ASF dual-hosted git repository.
lausen pushed a change to branch leezu-patch-2
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
discard 9fedcf4 Update test_profiler_gpu.py
discard b977b77 Update test_profiler_gpu.py
discard 1711d3e Mark GPU memory profiler tests as flaky
add ce37302 Numpy vstack large tensor fix (#19313)
add 94b649f BUGFIX Fix ConvTranspose __repr__ (#19338) (#19344)
add 93ddbbf Numpy diag large tensor fix (#19302)
add ce1e682 Add FindCUTENSOR.cmake (#19334)
add d0ceecb [DOC] Add shield for pr-awaiting-review and link to codereview page (#19352)
add 7bbe928 adding large tensor tests to verify support for split, hsplit, vsplit, dsplit (#19357)
add dfda45b fix seg fault (#19324)
add e6cea0d fix numpy ediff1d large tensor (#19310)
add a0fd1fe Faster pointwise fusion graph pass (#19269)
add defaafe Fix AtWithRecord (#19374)
add 3faf6df Numpy nan_to_num large tensor fix (#19350)
add 75c6216 Numpy Argmax Rewrite (#19339)
add 9e9f972 Move AMP from contrib to core (#19347)
add 43750c8 Remove cleanup on side threads (#19378)
add 3f833e1 Numpy histogram large tensor fix (#19392)
add 3f436fb Remove duplicate @with_seed decorators (#19336)
add eed080f argmin optimization (#19404)
add 187c75d Enable large tensor support for interp (#19363)
add 8e5a16c More extensions fixes (#19393)
add 3dda8e9 Enable large tensor support for padding (#19403)
add 9bfe311 Update elemwise_binary_broadcast_op_basic.cu (#18761)
add 329aebd initial disclaimer update (#19402) (#19416)
add 95f9ea2 fix (#19376)
add 4b3be14 Fix windows dll loading for compute capabilties >7.5 (#19410)
add 8dc3652 Add option to build with shared c runtime on windows (#19409)
add f9fef4d [LICENSE] Change intgemm to a submodule instead of fetch. (#19408)
add bd55002 fix (#19366)
add 383f8d1 fix fill_diagonal against large tensors (#19429)
add c6c5e4b Numpy diff large tensor fix (#19365)
add 0743fc2 Update oneDNN to v1.6.5 (#19444)
add 0514233 [TUTORIAL] Update crashcourse for MXNet 2 (#19345)
add 0faecf0 [BUGFIX] Fix MKLDNN BatchNorm with even number of channels (#19150) #19299 (#19425)
add cec6bcf operators for the sliding window self-attention (#19387)
add e486704 Mark test_foreach as garbage_expected (#19459)
add 33d94f1 Enable Numpy support for Gluon Block optimize_for (#19455)
add ea222a3 Update README.md (#19438)
add b33fbd1 Make RTC compatible with CUDA enhanced compatibility (#19364)
add 3d1df4e Support both Gluon 1 and 2 in the hybrid containers (#19470)
add aa4e3ac Mark test_lstm_clip as flaky (#19474)
add 07cd205 [FIX] CheckpointHandler verbose behavior in gluon.contrib.estimator.event_handler (#19471)
add 087f6ff [BUGFIX] Fallback to native implementation of RNN when use_sequence_length=True (#19466)
add 27ae77d numpy tril_indices large tensor fix (#19381)
add 94f55b1 Avoid segfault when calling __repr__ on ndarray or symbol with freed handle (#19475)
add ba96c06 Skip GPU memory profiler tests due to flaky segmentation faults
add bae10e6 Skip test_aggregate_duplication(
This update added new revisions after undoing existing revisions.
That is to say, some revisions that were in the old version of the
branch are not in the new version. This situation occurs
when a user --force pushes a change and generates a repository
containing something like this:
* -- * -- B -- O -- O -- O (9fedcf4)
\
N -- N -- N refs/heads/leezu-patch-2 (bae10e6)
You should already have received notification emails for all of the O
revisions, and so the following emails describe only the N revisions
from the common base, B.
Any revisions marked "omit" are not gone; other references still
refer to them. Any revisions marked "discard" are gone forever.
No new revisions were added by this update.
Summary of changes:
.gitmodules | 3 +
3rdparty/intgemm | 1 +
3rdparty/mkldnn | 2 +-
CMakeLists.txt | 114 ++--
DISCLAIMER-WIP | 7 +-
LICENSE | 37 ++
README.md | 4 +-
.../{FindCUDNN.cmake => FindCUTENSOR.cmake} | 20 +-
config/darwin.cmake | 1 +
config/linux.cmake | 1 +
config/linux_gpu.cmake | 1 +
conftest.py | 13 +-
.../getting-started/crash-course/0-introduction.md | 78 +++
.../getting-started/crash-course/1-ndarray.md | 121 -----
.../getting-started/crash-course/1-nparray.md | 211 ++++++++
.../getting-started/crash-course/2-create-nn.md | 532 ++++++++++++++++++
.../tutorials/getting-started/crash-course/2-nn.md | 150 ------
.../getting-started/crash-course/3-autograd.md | 214 +++++++-
.../getting-started/crash-course/4-components.md | 379 +++++++++++++
.../getting-started/crash-course/4-train.md | 178 ------
.../getting-started/crash-course/5-datasets.md | 310 +++++++++++
.../getting-started/crash-course/5-predict.md | 159 ------
.../getting-started/crash-course/6-train-nn.md | 442 +++++++++++++++
.../getting-started/crash-course/6-use_gpus.md | 151 ------
.../getting-started/crash-course/7-use-gpus.md | 253 +++++++++
.../getting-started/crash-course/index.rst | 22 +-
.../crash-course/prepare_dataset.py | 58 ++
.../python/tutorials/performance/backend/amp.md | 2 +-
.../static_site/src/pages/community/code_review.md | 5 +
.../amp_model_conversion.py | 2 +-
include/mxnet/lib_api.h | 12 +-
include/mxnet/libinfo.h | 1 +
include/mxnet/rtc.h | 2 +-
python/mxnet/_ctypes/ndarray.py | 4 +-
python/mxnet/_ctypes/symbol.py | 4 +-
python/mxnet/{contrib => }/amp/__init__.py | 0
python/mxnet/{contrib => }/amp/amp.py | 28 +-
python/mxnet/{contrib => }/amp/lists/__init__.py | 0
.../mxnet/{contrib => }/amp/lists/symbol_bf16.py | 0
.../mxnet/{contrib => }/amp/lists/symbol_fp16.py | 5 +-
python/mxnet/{contrib => }/amp/loss_scaler.py | 6 +-
python/mxnet/cython/ndarray.pyx | 3 +
python/mxnet/cython/symbol.pyx | 3 +
python/mxnet/gluon/block.py | 13 +-
.../mxnet/gluon/contrib/estimator/event_handler.py | 2 +-
python/mxnet/gluon/nn/basic_layers.py | 43 +-
python/mxnet/gluon/nn/conv_layers.py | 6 +-
python/mxnet/gluon/parameter.py | 9 +-
python/mxnet/ndarray/ndarray.py | 11 +-
python/mxnet/numpy/multiarray.py | 29 +-
python/mxnet/symbol/numpy/_symbol.py | 11 +-
python/mxnet/symbol/symbol.py | 13 +-
python/mxnet/test_utils.py | 5 +
src/api/operator/numpy/np_init_op.cc | 10 +-
src/api/operator/numpy/np_matrix_op.cc | 22 +-
src/c_api/c_api.cc | 26 +-
src/common/cuda/rtc.cc | 48 +-
src/common/rtc.cc | 37 +-
src/engine/naive_engine.cc | 3 -
src/engine/threaded_engine_perdevice.cc | 4 -
src/imperative/cached_op.h | 5 +-
src/imperative/exec_pass.h | 16 +-
src/imperative/pointwise_fusion_pass.cc | 517 ++++++++++--------
src/imperative/simple_partition_pass.cc | 265 +++++++++
src/imperative/simple_partition_pass.h | 599 +++++++--------------
src/libinfo.cc | 2 +
src/ndarray/ndarray.cc | 33 +-
src/operator/{contrib => }/all_finite-inl.h | 18 +-
src/operator/{contrib => }/all_finite.cc | 0
src/operator/{contrib => }/all_finite.cu | 0
src/operator/{contrib => }/amp_graph_pass.cc | 0
src/operator/contrib/transformer-inl.h | 222 ++++++++
src/operator/contrib/transformer.cc | 191 +++++++
src/operator/contrib/transformer.cu | 16 +
src/operator/fusion/fused_op.cu | 51 +-
src/operator/mshadow_op.h | 160 ++++++
src/operator/nn/mkldnn/mkldnn_batch_norm-inl.h | 16 +-
src/operator/nn/mkldnn/mkldnn_rnn-inl.h | 5 +
src/operator/nn/moments-inl.h | 12 +-
src/operator/numpy/np_broadcast_reduce_op.cuh | 44 ++
src/operator/numpy/np_broadcast_reduce_op.h | 98 +++-
src/operator/numpy/np_broadcast_reduce_op_index.cc | 12 +-
src/operator/numpy/np_broadcast_reduce_op_index.cu | 4 +-
src/operator/numpy/np_diff-inl.h | 20 +-
src/operator/numpy/np_ediff1d_op-inl.h | 2 +-
src/operator/numpy/np_einsum_op.cu | 29 +-
src/operator/numpy/np_fill_diagonal_op-inl.h | 2 +-
src/operator/numpy/np_init_op.h | 4 +-
src/operator/numpy/np_interp_op-inl.h | 28 +-
src/operator/numpy/np_kron-inl.h | 4 +-
src/operator/numpy/np_matrix_op-inl.h | 50 +-
src/operator/numpy/np_matrix_op.cc | 23 +-
src/operator/numpy/np_pad_op-inl.h | 72 +--
src/operator/numpy/np_tensordot_op-inl.h | 2 +-
src/operator/rnn.cc | 6 +-
src/operator/tensor/broadcast_reduce-inl.cuh | 28 +-
src/operator/tensor/broadcast_reduce-inl.h | 14 +-
.../tensor/elemwise_binary_broadcast_op_basic.cu | 2 +-
src/operator/tensor/elemwise_unary_op.h | 4 +-
src/operator/tensor/histogram.cc | 4 +-
src/operator/tensor/init_op.h | 4 +-
.../nightly/apache_rat_license_check/rat-excludes | 1 +
tests/nightly/common.py | 100 ++++
tests/nightly/estimator/test_estimator_cnn.py | 7 +-
tests/nightly/estimator/test_sentiment_rnn.py | 7 +-
tests/nightly/test_large_array.py | 3 -
tests/nightly/test_large_vector.py | 3 -
tests/nightly/test_np_large_array.py | 396 +++++++++++++-
tests/nightly/test_np_random.py | 11 +-
.../gpu/{test_contrib_amp.py => test_amp.py} | 9 +-
tests/python/gpu/test_fusion.py | 25 +-
tests/python/gpu/test_gluon_gpu.py | 21 +-
tests/python/gpu/test_gluon_model_zoo_gpu.py | 3 -
tests/python/gpu/test_gluon_transforms.py | 6 +-
tests/python/gpu/test_kvstore_gpu.py | 2 -
tests/python/gpu/test_numpy_fallback.py | 3 +-
tests/python/gpu/test_numpy_op.py | 3 +-
tests/python/gpu/test_operator_gpu.py | 61 +--
.../mkl/{test_contrib_amp.py => test_amp.py} | 7 +-
tests/python/mkl/test_bf16_operator.py | 15 +-
tests/python/mkl/test_mkldnn.py | 26 +-
tests/python/mkl/test_subgraph.py | 1 -
tests/python/unittest/common.py | 76 ---
tests/python/unittest/onnx/mxnet_export_test.py | 7 -
tests/python/unittest/test_autograd.py | 20 +-
tests/python/unittest/test_contrib_control_flow.py | 13 +-
.../unittest/test_contrib_gluon_data_vision.py | 4 -
tests/python/unittest/test_contrib_intgemm.py | 5 -
tests/python/unittest/test_contrib_operator.py | 5 +-
tests/python/unittest/test_contrib_optimizer.py | 3 +-
tests/python/unittest/test_contrib_stes_op.py | 4 +-
tests/python/unittest/test_deferred_compute.py | 1 +
tests/python/unittest/test_dynamic_shape.py | 1 -
tests/python/unittest/test_exc_handling.py | 12 -
tests/python/unittest/test_executor.py | 5 -
tests/python/unittest/test_gluon.py | 225 ++++----
tests/python/unittest/test_gluon_data.py | 18 -
tests/python/unittest/test_gluon_data_vision.py | 21 +-
tests/python/unittest/test_gluon_model_zoo.py | 3 -
tests/python/unittest/test_gluon_probability_v1.py | 35 --
tests/python/unittest/test_gluon_probability_v2.py | 34 --
tests/python/unittest/test_gluon_rnn.py | 16 +-
tests/python/unittest/test_gluon_trainer.py | 13 +-
tests/python/unittest/test_higher_order_grad.py | 33 +-
tests/python/unittest/test_image.py | 18 +-
tests/python/unittest/test_infer_type.py | 4 +-
tests/python/unittest/test_kvstore.py | 12 +-
tests/python/unittest/test_kvstore_custom.py | 9 +-
tests/python/unittest/test_loss.py | 7 +-
tests/python/unittest/test_metric.py | 2 +-
tests/python/unittest/test_ndarray.py | 57 +-
.../test_numpy_contrib_gluon_data_vision.py | 5 -
tests/python/unittest/test_numpy_default_dtype.py | 2 -
tests/python/unittest/test_numpy_gluon.py | 39 +-
.../unittest/test_numpy_gluon_data_vision.py | 22 +-
.../python/unittest/test_numpy_interoperability.py | 6 +-
tests/python/unittest/test_numpy_loss.py | 7 +-
tests/python/unittest/test_numpy_ndarray.py | 25 +-
tests/python/unittest/test_numpy_op.py | 186 +------
tests/python/unittest/test_operator.py | 254 +++------
tests/python/unittest/test_optimizer.py | 25 +-
tests/python/unittest/test_profiler.py | 1 +
tests/python/unittest/test_random.py | 26 +-
tests/python/unittest/test_recordio.py | 4 -
tests/python/unittest/test_sparse_ndarray.py | 41 +-
tests/python/unittest/test_sparse_operator.py | 25 +-
tests/python/unittest/test_subgraph.py | 2 -
tests/python/unittest/test_tvm_op.py | 2 -
tools/pip/setup.py | 4 +
tools/windowsbuild/warp_dll.cpp | 7 +-
170 files changed, 5406 insertions(+), 3065 deletions(-)
create mode 160000 3rdparty/intgemm
copy cmake/Modules/{FindCUDNN.cmake => FindCUTENSOR.cmake} (57%)
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/0-introduction.md
delete mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/1-ndarray.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/1-nparray.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/2-create-nn.md
delete mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/2-nn.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/4-components.md
delete mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/4-train.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/5-datasets.md
delete mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/5-predict.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/6-train-nn.md
delete mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/6-use_gpus.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/7-use-gpus.md
create mode 100644 docs/python_docs/python/tutorials/getting-started/crash-course/prepare_dataset.py
rename python/mxnet/{contrib => }/amp/__init__.py (100%)
rename python/mxnet/{contrib => }/amp/amp.py (98%)
rename python/mxnet/{contrib => }/amp/lists/__init__.py (100%)
rename python/mxnet/{contrib => }/amp/lists/symbol_bf16.py (100%)
rename python/mxnet/{contrib => }/amp/lists/symbol_fp16.py (99%)
rename python/mxnet/{contrib => }/amp/loss_scaler.py (96%)
create mode 100644 src/imperative/simple_partition_pass.cc
rename src/operator/{contrib => }/all_finite-inl.h (88%)
mode change 100755 => 100644
rename src/operator/{contrib => }/all_finite.cc (100%)
mode change 100755 => 100644
rename src/operator/{contrib => }/all_finite.cu (100%)
mode change 100755 => 100644
rename src/operator/{contrib => }/amp_graph_pass.cc (100%)
create mode 100644 src/operator/numpy/np_broadcast_reduce_op.cuh
create mode 100644 tests/nightly/common.py
rename tests/python/gpu/{test_contrib_amp.py => test_amp.py} (97%)
rename tests/python/mkl/{test_contrib_amp.py => test_amp.py} (97%)