Skip to content

N
ngraph
Commit 6ca1a511 authored by amy.zhuang's avatar amy.zhuang
Browse files
Options

Merge branch 'master' into ayzhuang/in-place-concat

parents 749a1c6b 9debc0bc
Expand all Hide whitespace changes
Inline Side-by-side
Showing with 1267 additions and 928 deletions
.ci/onnx/jenkins/prepare_environment.sh
View file @ 6ca1a511
......@@ -68,7 +68,7 @@ function build_ngraph() {
make install || return 1
cd "${ngraph_directory}/ngraph/python"
if [ ! -d ./pybind11 ]; then
git clone --recursive -b allow-nonconstructible-holders https://github.com/jagerman/pybind11.git
git clone --recursive https://github.com/pybind/pybind11.git
fi
export PYBIND_HEADERS_PATH="${ngraph_directory}/ngraph/python/pybind11"
export NGRAPH_CPP_BUILD_PATH="${ngraph_directory}/ngraph_dist"
......
.ci/travis/ubuntu/Dockerfile
View file @ 6ca1a511
......@@ -39,7 +39,7 @@ RUN make install
# Prepare nGraph Python API
WORKDIR /root/ngraph/python
RUN git clone --recursive -b allow-nonconstructible-holders https://github.com/jagerman/pybind11.git
RUN git clone --recursive https://github.com/pybind/pybind11.git
ENV NGRAPH_CPP_BUILD_PATH /root/ngraph_dist
ENV LD_LIBRARY_PATH /root/ngraph_dist/lib
ENV PYBIND_HEADERS_PATH /root/ngraph/python/pybind11
......
CMakeLists.txt
View file @ 6ca1a511
......@@ -196,25 +196,8 @@ if(WIN32)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -D_CRT_SECURE_NO_WARNINGS")
endif()
include(unit_test_control)
set(UNIT_TEST_CONFIG_LIST "" CACHE INTERNAL "")
if (NGRAPH_INTERPRETER_ENABLE)
unit_test_control(BACKEND INTERPRETER MANIFEST src/ngraph/runtime/interpreter/unit_test.manifest)
endif()
# Set true if CPU backend is built by default
if (NGRAPH_CPU_ENABLE)
unit_test_control(BACKEND CPU MANIFEST src/ngraph/runtime/cpu/unit_test.manifest)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DNGRAPH_CPU_ENABLE")
endif()
if (NGRAPH_INTELGPU_ENABLE)
unit_test_control(BACKEND INTELGPU MANIFEST src/ngraph/runtime/intelgpu/unit_test.manifest)
endif()
if (NGRAPH_GPU_ENABLE)
unit_test_control(BACKEND GPU MANIFEST src/ngraph/runtime/gpu/unit_test.manifest)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DNGRAPH_CPU_ENABLE")
endif()
if (NOT DEFINED NGRAPH_TBB_ENABLE)
......
README.md
View file @ 6ca1a511
# nGraph Library [![Build Status][build-status-badge]][build-status]
Welcome to the open-source repository for the Intel® nGraph™ Library. Our code
Welcome to the open-source repository for the **Intel® nGraph Library**. Our code
base provides a Compiler and runtime suite of tools (APIs) designed to give
developers maximum flexibility for their software design, allowing them to
create or customize a scalable solution using any framework while also avoiding
......@@ -11,7 +10,9 @@ backends, and it will be able to run on any backends we support in the future
with minimal disruption to your model. With nGraph, you can co-evolve your
software and hardware's capabilities to stay at the forefront of your industry.
The nGraph Compiler is Intel's graph compiler for Artificial Neural Networks.
![nGraph ecosystem][ngraph-ecosystem]
The **nGraph Compiler** is Intel's graph compiler for Artificial Neural Networks.
Documentation in this repo describes how you can program any framework
to run training and inference computations on a variety of Backends including
Intel® Architecture Processors (CPUs), Intel® Nervana™ Neural Network Processors
......@@ -24,18 +25,29 @@ whatever scenario you need.
nGraph provides both a C++ API for framework developers and a Python API which
can run inference on models imported from ONNX.
![nGraph ecosystem][ngraph-ecosystem]
See the [Release Notes] for recent changes.
| Framework | bridge available? | ONNX support? |
|----------------|-------------------|----------------|
| TensorFlow* | yes | yes |
| MXNet* | yes | yes |
| PaddlePaddle | yes | yes |
| PyTorch* | no | yes |
| Chainer* | no | yes |
| CNTK* | no | yes |
| Caffe2* | no | yes |
|Framework | bridge available? | ONNX support? |
|------------|-------------------|----------------|
| neon | yes | yes |
| MXNet* | yes | yes |
| TensorFlow*| yes | yes |
| PyTorch* | not yet | yes |
| Chainer* | not yet | yes |
| CNTK* | not yet | yes |
| Caffe2* | not yet | yes |
| Backend | current support | future support |
|-----------------------------------------------|-------------------|----------------|
| Intel® Architecture CPU | yes | yes |
| Intel® Nervana™ Neural Network Processor (NNP)| yes | yes |
| Intel [Movidius™ Myriad™ 2] VPUs | coming soon | yes |
| Intel® Architecture GPUs | via PlaidML | yes |
| AMD* GPUs | via PlaidML | yes |
| NVIDIA* GPUs | via PlaidML | some |
| Field Programmable Gate Arrays (FPGA) | no | yes |
## Documentation
......@@ -72,12 +84,13 @@ to improve the Library:
[install]: http://ngraph.nervanasys.com/docs/latest/buildlb.html
[framework integration guides]: http://ngraph.nervanasys.com/docs/latest/framework-integration-guides.html
[release notes]: http://ngraph.nervanasys.com/docs/latest/project/release-notes.html
[Github issues]: https://github.com/NervanaSystems/ngraph/issues
[contrib guide]: http://ngraph.nervanasys.com/docs/latest/project/code-contributor-README.html
[pull request]: https://github.com/NervanaSystems/ngraph/pulls
[how to import]: http://ngraph.nervanasys.com/docs/latest/howto/import.html
[ngraph-ecosystem]: doc/sphinx/source/graphics/ngraph-ecosystem.png "nGraph Ecosystem"
[ngraph-ecosystem]: doc/sphinx/source/graphics/599px-Intel-ngraph-ecosystem.png "nGraph Ecosystem"
[build-status]: https://travis-ci.org/NervanaSystems/ngraph/branches
[build-status-badge]: https://travis-ci.org/NervanaSystems/ngraph.svg?branch=master
[develop-without-lockin]: doc/sphinx/source/graphics/develop-without-lockin.png "Develop on any part of the stack wtihout lockin"
[Movidius]:https://www.movidius.com/solutions/vision-processing-unit
[Movidius™ Myriad™ 2]:https://www.movidius.com/solutions/vision-processing-unit
cmake/Modules/unit_test_control.cmake deleted 100644 → 0
View file @ 749a1c6b
# ******************************************************************************
# Copyright 2017-2018 Intel Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ******************************************************************************
function(UNIT_TEST_CONTROL)
set(options)
set(oneValueArgs BACKEND MANIFEST)
set(multiValueArgs)
cmake_parse_arguments(UNIT_TEST_CONTROL "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
if (UNIT_TEST_CONTROL_MANIFEST)
get_filename_component(UNIT_TEST_CONTROL_MANIFEST ${UNIT_TEST_CONTROL_MANIFEST} ABSOLUTE)
set(CONFIG_STRING "${UNIT_TEST_CONTROL_BACKEND}@${UNIT_TEST_CONTROL_MANIFEST}")
else()
set(CONFIG_STRING "${UNIT_TEST_CONTROL_BACKEND}@")
endif()
set(UNIT_TEST_CONFIG_LIST "${UNIT_TEST_CONFIG_LIST};${CONFIG_STRING}" CACHE INTERNAL "")
endfunction()
doc/sphinx/source/conf.py
View file @ 6ca1a511
......@@ -68,9 +68,9 @@ author = 'Intel Corporation'
# built documents.
#
# The short X.Y version.
version = '0.8'
version = '0.9'
# The full version, including alpha/beta/rc tags.
release = '0.8.0'
release = '0.9.0'
# The language for content autogenerated by Sphinx. Refer to documentation
# for a list of supported languages.
......
doc/sphinx/source/index.rst
View file @ 6ca1a511
......@@ -30,8 +30,8 @@ software engineers, and others with the means to make their work :ref:`portable`
:abbr:`Machine Learning (ML)` hardware available today: optimized Deep Learning
computation devices.
.. figure:: graphics/ngraph-ecosystem.png
:width: 650px
.. figure:: graphics/599px-Intel-ngraph-ecosystem.png
:width: 599px
.. _portable:
......@@ -66,16 +66,15 @@ Python-based API. See the `ngraph onnx companion tool`_ to get started.
.. csv-table::
:header: "Framework", "Bridge Code Available?", "ONNX Support?"
:header: "Framework", "Bridge Available?", "ONNX Support?"
:widths: 27, 10, 10
TensorFlow, Yes, Yes
MXNet, Yes, Yes
PaddlePaddle, Coming Soon, Yes
neon, none needed, Yes
PyTorch, Coming Soon, Yes
CNTK, Not yet, Yes
Other, Not yet, Doable
PyTorch, No, Yes
CNTK, No, Yes
Other, Custom, Custom
.. _deployable:
......@@ -104,29 +103,30 @@ model to run on a variety of backends:
.. csv-table::
:header: "Backend", "Current nGraph support", "Future nGraph support"
:header: "Backend", "Current support", "Future nGraph support"
:widths: 35, 10, 10
Intel® Architecture Processors (CPUs), Yes, Yes
Intel® Nervana™ Neural Network Processor™ (NNPs), Yes, Yes
NVIDIA\* CUDA (GPUs), Yes, Some
Intel® Nervana™ Neural Network Processor (NNPs), Yes, Yes
AMD\* GPUs, via PlaidML, Yes
NVIDIA\* GPUs, via PlaidML, Some
Intel® Architecture GPUs, Yes, Yes
:abbr:`Field Programmable Gate Arrays (FPGA)` (FPGAs), Coming soon, Yes
`Movidius`_, Not yet, Yes
Intel Movidius™ Myriad™ 2 (VPU), Coming soon, Yes
Other, Not yet, Ask
The value we're offering to the developer community is empowerment: we are
confident that Intel® Architecture already provides the best computational
resources available for the breadth of ML/DL tasks. We welcome ideas and
`contributions`_ from the community.
resources available for the breadth of ML/DL tasks. We welcome ideas and
`contributions`_ from the community.
Further project details can be found on our :doc:`project/about` page, or see
our :doc:`buildlb` guide for how to get started.
our :doc:`buildlb` guide for how to get started.
.. note:: The library code is under active development as we're continually
.. note:: The Library code is under active development as we're continually
adding support for more kinds of DL models and ops, framework compiler
optimizations, and backends.
optimizations, and backends.
=======
......@@ -152,7 +152,6 @@ Contents
project/index.rst
Indices and tables
==================
......@@ -160,8 +159,7 @@ Indices and tables
* :ref:`genindex`
.. _ONNX: http://onnx.ai
.. _ONNX: http://onnx.ai
.. _ngraph onnx companion tool: https://github.com/NervanaSystems/ngraph-onnx
.. _Movidius: https://www.movidius.com/
.. _contributions: https://github.com/NervanaSystems/ngraph#how-to-contribute
.. _contributions: https://github.com/NervanaSystems/ngraph#how-to-contribute
\ No newline at end of file
doc/sphinx/source/project/about.rst
View file @ 6ca1a511
......@@ -5,7 +5,7 @@ Overview
Welcome to the documentation site for |InG|, an open-source C++ Compiler,
Library, and runtime suite for running training and inference on
Library, and runtime suite for Deep Learning frameworks running training and inference on
:abbr:`Deep Neural Network (DNN)` models. nGraph is framework-neutral and can be
targeted for programming and deploying :abbr:`Deep Learning (DL)` applications
on the most modern compute and edge devices.
......@@ -22,8 +22,8 @@ Features
Develop without lock-in
~~~~~~~~~~~~~~~~~~~~~~~
.. figure:: ../graphics/develop-without-lockin.png
:width: 650px
.. figure:: ../graphics/599px-Intel-ngraph-ecosystem.png
:width: 599px
Being able to increase training performance or reduce inference latency by simply
......@@ -34,47 +34,6 @@ developers working with nGraph. Our commitment to bake flexibility into our
ecosystem ensures developers' freedom to design user-facing APIs for various
hardware deployments directly into their frameworks.
.. figure:: ../graphics/ngraph-ecosystem.png
:width: 585px
nGraph currently supports :doc:`three popular <../framework-integration-guides>`
frameworks for :abbr:`Deep Learning (DL)` models through what we call
a :term:`bridge` that can be integrated during the framework's build time.
For developers working with other frameworks (even those not listed above),
we've created a :doc:`How to Guide <../howto/index>` so you can learn how to
create custom bridge code that can be used to
:doc:`compile and run <../howto/execute>` a training model.
Additionally, nGraph Library supports the `ONNX`_ format. Developers who
already have a "trained" model can use nGraph to bypass much of the
framework-based complexity and :doc:`../howto/import` to test or run it
on targeted and efficient backends with our user-friendly ``ngraph_api``.
With nGraph, data scientists can focus on data science rather than worrying
about how to adapt models to train and run efficiently on different devices.
Be sure to add the ``-DNGRAPH_ONNX_IMPORT_ENABLE=ON`` option when running `cmake`
to build the Library.
Supported platforms
--------------------
* Intel® Architecture Processors (CPUs),
* Intel® Nervana™ Neural Network Processor™ (NNPs), and
* NVIDIA\* CUDA (GPUs).
We built the first-generation of the Intel Nervana™ NNP family of processors
last year to show that the nGraph Library can be used to train a
:abbr:`Neural Network (NN)` more quickly. The more advanced the silicon, the
more powerful a lightweight a library can be. So while we do currently support
traditional GPUs, they are not advanced silicon, and trying to scale workloads
using traditional GPU libraries is clunky and brittle with bottlenecks. Iteration
from an already-trained NN model to one that can also perform inference
computations is immensely simplified. Read more about these compute-friendly
options on the documentation for :doc:`../fusion/index`.
.. note:: The library code is under active development as we're continually
adding support for more kinds of DL models and ops, framework compiler
optimizations, and backends.
......
doc/sphinx/source/project/doc-contributor-README.rst
View file @ 6ca1a511
......@@ -15,13 +15,13 @@
.. limitations under the License.
.. ---------------------------------------------------------------------------
nGraph™ Library docs
====================
nGraph Library docs
===================
Read this for changes affecting anything in ``ngraph/doc``
----------------------------------------------------------
For updates to the Intel® nGraph™ Library ``/doc`` repo, please submit a PR with
For updates to the Intel® nGraph Library ``/doc`` repo, please submit a PR with
any changes or ideas you'd like integrated. This helps us maintain trackability
with respect to additions or feature requests.
......@@ -127,7 +127,7 @@ To build documentation locally, run:
.. code-block:: console
$ pip3 install [-I] Sphinx==1.6.5 [--user]
$ pip3 install [-I] Sphinx==1.7.5 [--user]
$ pip3 install [-I] breathe numpy [--user]
$ cd doc/sphinx/
$ make html
......@@ -150,11 +150,16 @@ To build documentation in a python3 virtualenv, run:
Then point your browser at ``localhost:8000``.
.. note:: For docs built in a virtual env, Sphinx latest changes may break
documentation; try building with a specific version of Sphinx.
For tips on writing reStructuredText-formatted documentation, see the `sphinx`_
stable reST documentation.
.. _ngraph repo: https://github.com/NervanaSystems/ngraph-cpp/
.. _documentation repo: https://github.com/NervanaSystems/private-ngraph/tree/master/doc
.. _ngraph repo: https://github.com/NervanaSystems/ngraph/
.. _documentation repo: https://github.com/NervanaSystems/ngraph/tree/master/doc
.. _sphinx: http://www.sphinx-doc.org/en/stable/rest.html
.. _wiki: https://github.com/NervanaSystems/ngraph/wiki/
.. _breathe: https://breathe.readthedocs.io/en/latest/
......
doc/sphinx/source/project/release-notes.rst
View file @ 6ca1a511
......@@ -3,5 +3,12 @@
Release Notes
#############
This is the |version| of release.
This is release |release|.
API Changes
===========
.. literalinclude:: ../../../../changes.md
src/ngraph/frontend/onnx_import/op/abs.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector abs(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Abs>(node.get_ng_inputs().at(0))};
}
inline NodeVector abs(const Node& node)
{
return {std::make_shared<ngraph::op::Abs>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/add.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector add(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Add>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector add(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Add>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/and.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector logical_and(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::And>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector logical_and(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::And>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/average_pool.cpp
View file @ 6ca1a511
......@@ -26,12 +26,16 @@ namespace ngraph
{
namespace op
{
NodeVector average_pool(const Node& node)
namespace set_1
{
return convpool::make_ng_pool<ngraph::op::AvgPool>(node);
}
NodeVector average_pool(const Node& node)
{
return convpool::make_ng_pool<ngraph::op::AvgPool>(node);
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/average_pool.hpp
View file @ 6ca1a511
......@@ -26,17 +26,21 @@ namespace ngraph
{
namespace op
{
/**
* @brief Convert ONNX AveragePool operation to an nGraph node.
*
* @param node The ONNX node object representing this operation.
*
* @return The vector containing Ngraph nodes producing output of ONNX AveragePool
* operation.
*/
NodeVector average_pool(const Node& node);
} // namespace op
namespace set_1
{
/**
* @brief Convert ONNX AveragePool operation to an nGraph node.
*
* @param node The ONNX node object representing this operation.
*
* @return The vector containing Ngraph nodes producing output of ONNX AveragePool
* operation.
*/
NodeVector average_pool(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/batch_norm.cpp
View file @ 6ca1a511
......@@ -28,38 +28,42 @@ namespace ngraph
{
namespace op
{
NodeVector batch_norm(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto x = inputs.at(0);
auto scale = inputs.at(1);
auto bias = inputs.at(2);
std::shared_ptr<ngraph::Node> mean{nullptr};
std::shared_ptr<ngraph::Node> var{nullptr};
NodeVector batch_norm(const Node& node)
{
NodeVector inputs{node.get_ng_inputs()};
auto x = inputs.at(0);
auto scale = inputs.at(1);
auto bias = inputs.at(2);
std::shared_ptr<ngraph::Node> mean{nullptr};
std::shared_ptr<ngraph::Node> var{nullptr};
int is_test{node.get_attribute_value<int>("is_test", 1)};
int spatial{node.get_attribute_value<int>("spatial", 1)};
double epsilon{node.get_attribute_value<double>("epsilon", 1e-5)};
// TODO: Implement learning mode support
// float momentum{node.get_attribute_value<float>("momentum", 0.9f)};
bool training = false;
int is_test{node.get_attribute_value<int>("is_test", 1)};
int spatial{node.get_attribute_value<int>("spatial", 1)};
double epsilon{node.get_attribute_value<double>("epsilon", 1e-5)};
// TODO: Implement learning mode support
// float momentum{node.get_attribute_value<float>("momentum", 0.9f)};
bool training = false;
ASSERT_IS_SUPPORTED(node, is_test) << "only 'is_test' mode is supported.";
ASSERT_IS_SUPPORTED(node, spatial) << "only 'spatial' mode is supported.";
ASSERT_IS_SUPPORTED(node, is_test) << "only 'is_test' mode is supported.";
ASSERT_IS_SUPPORTED(node, spatial) << "only 'spatial' mode is supported.";
if (inputs.size() >= 5)
{
mean = inputs.at(3);
var = inputs.at(4);
return {std::make_shared<ngraph::op::BatchNorm>(
epsilon, scale, bias, x, mean, var, training)};
if (inputs.size() >= 5)
{
mean = inputs.at(3);
var = inputs.at(4);
return {std::make_shared<ngraph::op::BatchNorm>(
epsilon, scale, bias, x, mean, var, training)};
}
return {std::make_shared<ngraph::op::BatchNorm>(epsilon, scale, bias, x)};
}
return {std::make_shared<ngraph::op::BatchNorm>(epsilon, scale, bias, x)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
} // namespace onnx_import
} // namespace ngraph
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/batch_norm.hpp
View file @ 6ca1a511
......@@ -24,9 +24,14 @@ namespace ngraph
{
namespace op
{
NodeVector batch_norm(const Node& node);
} // namespace op
namespace set_1
{
NodeVector batch_norm(const Node& node);
} // namespace onnx_import
} // namespace set_1
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/cast.cpp
View file @ 6ca1a511
......@@ -30,34 +30,40 @@ namespace ngraph
{
namespace op
{
NodeVector cast(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
int64_t target_type = node.get_attribute_value<int64_t>("to");
element::Type elem_type;
switch (target_type)
NodeVector cast(const Node& node)
{
case onnx::TensorProto_DataType_BOOL: elem_type = element::boolean; break;
case onnx::TensorProto_DataType_DOUBLE: elem_type = element::f64; break;
case onnx::TensorProto_DataType_FLOAT16:
case onnx::TensorProto_DataType_FLOAT: elem_type = element::f32; break;
case onnx::TensorProto_DataType_INT8: elem_type = element::i8; break;
case onnx::TensorProto_DataType_INT16: elem_type = element::i16; break;
case onnx::TensorProto_DataType_INT32: elem_type = element::i32; break;
case onnx::TensorProto_DataType_INT64: elem_type = element::i64; break;
case onnx::TensorProto_DataType_UINT8: elem_type = element::u8; break;
case onnx::TensorProto_DataType_UINT16: elem_type = element::u16; break;
case onnx::TensorProto_DataType_UINT32: elem_type = element::u32; break;
case onnx::TensorProto_DataType_UINT64: elem_type = element::u64; break;
case onnx::TensorProto_DataType_UNDEFINED: elem_type = element::unspecified; break;
default: ASSERT_IS_SUPPORTED(node, false) << "unsupported type";
auto data = node.get_ng_inputs().at(0);
int64_t target_type = node.get_attribute_value<int64_t>("to");
element::Type elem_type;
switch (target_type)
{
case onnx::TensorProto_DataType_BOOL: elem_type = element::boolean; break;
case onnx::TensorProto_DataType_DOUBLE: elem_type = element::f64; break;
case onnx::TensorProto_DataType_FLOAT16:
case onnx::TensorProto_DataType_FLOAT: elem_type = element::f32; break;
case onnx::TensorProto_DataType_INT8: elem_type = element::i8; break;
case onnx::TensorProto_DataType_INT16: elem_type = element::i16; break;
case onnx::TensorProto_DataType_INT32: elem_type = element::i32; break;
case onnx::TensorProto_DataType_INT64: elem_type = element::i64; break;
case onnx::TensorProto_DataType_UINT8: elem_type = element::u8; break;
case onnx::TensorProto_DataType_UINT16: elem_type = element::u16; break;
case onnx::TensorProto_DataType_UINT32: elem_type = element::u32; break;
case onnx::TensorProto_DataType_UINT64: elem_type = element::u64; break;
case onnx::TensorProto_DataType_UNDEFINED:
elem_type = element::unspecified;
break;
default: ASSERT_IS_SUPPORTED(node, false) << "unsupported type";
}
return {std::make_shared<ngraph::op::Convert>(data, elem_type)};
}
return {std::make_shared<ngraph::op::Convert>(data, elem_type)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/cast.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector cast(const Node& node);
} // namespace op
namespace set_1
{
NodeVector cast(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/ceil.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector ceil(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Ceiling>(node.get_ng_inputs().at(0))};
}
inline NodeVector ceil(const Node& node)
{
return {std::make_shared<ngraph::op::Ceiling>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/clip.cpp
View file @ 6ca1a511
......@@ -34,30 +34,37 @@ namespace ngraph
{
namespace op
{
NodeVector clip(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
NodeVector clip(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double max_value =
node.get_attribute_value<double>("max", std::numeric_limits<double>::max());
double min_value =
node.get_attribute_value<double>("min", std::numeric_limits<double>::lowest());
double max_value =
node.get_attribute_value<double>("max", std::numeric_limits<double>::max());
double min_value = node.get_attribute_value<double>(
"min", std::numeric_limits<double>::lowest());
std::shared_ptr<ngraph::Node> max_value_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{max_value});
max_value_node = make_broadcast_node(max_value_node, data->get_shape());
std::shared_ptr<ngraph::Node> max_value_node =
std::make_shared<ngraph::op::Constant>(data->get_element_type(),
ngraph::Shape{},
std::vector<double>{max_value});
max_value_node = make_broadcast_node(max_value_node, data->get_shape());
std::shared_ptr<ngraph::Node> min_value_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{min_value});
min_value_node = make_broadcast_node(min_value_node, data->get_shape());
std::shared_ptr<ngraph::Node> min_value_node =
std::make_shared<ngraph::op::Constant>(data->get_element_type(),
ngraph::Shape{},
std::vector<double>{min_value});
min_value_node = make_broadcast_node(min_value_node, data->get_shape());
return {std::make_shared<ngraph::op::Minimum>(
max_value_node, std::make_shared<ngraph::op::Maximum>(data, min_value_node))};
}
return {std::make_shared<ngraph::op::Minimum>(
max_value_node,
std::make_shared<ngraph::op::Maximum>(data, min_value_node))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/clip.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector clip(const Node& node);
} // namespace op
namespace set_1
{
NodeVector clip(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/concat.cpp
View file @ 6ca1a511
......@@ -24,16 +24,20 @@ namespace ngraph
{
namespace op
{
NodeVector concat(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto axis = node.get_attribute_value<int64_t>("axis");
NodeVector concat(const Node& node)
{
NodeVector inputs{node.get_ng_inputs()};
auto axis = node.get_attribute_value<int64_t>("axis");
return {std::make_shared<ngraph::op::Concat>(inputs, axis)};
}
return {std::make_shared<ngraph::op::Concat>(inputs, axis)};
}
} // namespace op
} // namespace set_1
} // namespace onnx_import
} //namespace op
} // namespace ngraph
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/concat.hpp
View file @ 6ca1a511
......@@ -26,9 +26,14 @@ namespace ngraph
{
namespace op
{
NodeVector concat(const Node& node);
} // namespace op
namespace set_1
{
NodeVector concat(const Node& node);
} // namespace onnx_import
} // namespace set_1
} // namespace ngraph
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/constant.cpp
View file @ 6ca1a511
......@@ -25,96 +25,101 @@ namespace ngraph
{
namespace op
{
namespace
namespace set_1
{
template <typename T>
inline std::shared_ptr<ngraph::op::Constant>
__make_ng_constant(const element::Type& type, const Tensor& tensor)
namespace
{
return std::make_shared<ngraph::op::Constant>(
type, tensor.get_shape(), tensor.get_data<T>());
}
template <typename T>
inline std::shared_ptr<ngraph::op::Constant>
__make_ng_constant(const element::Type& type, const Tensor& tensor)
{
return std::make_shared<ngraph::op::Constant>(
type, tensor.get_shape(), tensor.get_data<T>());
}
template <Tensor::Type>
inline std::shared_ptr<ngraph::op::Constant> make_ng_constant(const Tensor& tensor)
{
throw error::tensor::unsupported_data_type{tensor};
}
template <Tensor::Type>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant(const Tensor& tensor)
{
throw error::tensor::unsupported_data_type{tensor};
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float16>(const Tensor& tensor)
{
return __make_ng_constant<float>(element::f32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float16>(const Tensor& tensor)
{
return __make_ng_constant<float>(element::f32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float32>(const Tensor& tensor)
{
return __make_ng_constant<float>(element::f32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float32>(const Tensor& tensor)
{
return __make_ng_constant<float>(element::f32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float64>(const Tensor& tensor)
{
return __make_ng_constant<double>(element::f64, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::float64>(const Tensor& tensor)
{
return __make_ng_constant<double>(element::f64, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::int32>(const Tensor& tensor)
{
return __make_ng_constant<int32_t>(element::i32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::int32>(const Tensor& tensor)
{
return __make_ng_constant<int32_t>(element::i32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::int64>(const Tensor& tensor)
{
return __make_ng_constant<int64_t>(element::i64, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::int64>(const Tensor& tensor)
{
return __make_ng_constant<int64_t>(element::i64, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::uint32>(const Tensor& tensor)
{
return __make_ng_constant<uint32_t>(element::u32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::uint32>(const Tensor& tensor)
{
return __make_ng_constant<uint32_t>(element::u32, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::uint64>(const Tensor& tensor)
{
return __make_ng_constant<uint64_t>(element::u64, tensor);
}
template <>
inline std::shared_ptr<ngraph::op::Constant>
make_ng_constant<Tensor::Type::uint64>(const Tensor& tensor)
{
return __make_ng_constant<uint64_t>(element::u64, tensor);
}
inline std::shared_ptr<ngraph::op::Constant> make_constant(const Tensor& tensor)
{
inline std::shared_ptr<ngraph::op::Constant> make_constant(const Tensor& tensor)
{
#define MAKE_NG_CONSTANT(data_type_) \
case data_type_: return make_ng_constant<data_type_>(tensor)
switch (tensor.get_type())
{
MAKE_NG_CONSTANT(Tensor::Type::float16);
MAKE_NG_CONSTANT(Tensor::Type::float32);
MAKE_NG_CONSTANT(Tensor::Type::float64);
MAKE_NG_CONSTANT(Tensor::Type::int32);
MAKE_NG_CONSTANT(Tensor::Type::int64);
MAKE_NG_CONSTANT(Tensor::Type::uint32);
MAKE_NG_CONSTANT(Tensor::Type::uint64);
default: throw error::tensor::invalid_data_type{tensor};
switch (tensor.get_type())
{
MAKE_NG_CONSTANT(Tensor::Type::float16);
MAKE_NG_CONSTANT(Tensor::Type::float32);
MAKE_NG_CONSTANT(Tensor::Type::float64);
MAKE_NG_CONSTANT(Tensor::Type::int32);
MAKE_NG_CONSTANT(Tensor::Type::int64);
MAKE_NG_CONSTANT(Tensor::Type::uint32);
MAKE_NG_CONSTANT(Tensor::Type::uint64);
default: throw error::tensor::invalid_data_type{tensor};
}
}
}
}
NodeVector constant(const onnx_import::Node& node)
{
return {make_constant(node.get_attribute_value<Tensor>("value"))};
}
NodeVector constant(const onnx_import::Node& node)
{
return {make_constant(node.get_attribute_value<Tensor>("value"))};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/constant.hpp
View file @ 6ca1a511
......@@ -26,9 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector constant(const Node& node);
namespace set_1
{
NodeVector constant(const Node& node);
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/conv.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/frontend/onnx_import/op/conv.hpp
View file @ 6ca1a511
......@@ -26,15 +26,19 @@ namespace ngraph
{
namespace op
{
/// \brief Performs ONNX Conv operation.
///
/// \param node The ONNX node object representing this operation.
///
/// \return The vector containing Ngraph nodes producing output of ONNX convolution
/// operation.
NodeVector conv(const Node& node);
} // namespace op
namespace set_1
{
/// \brief Performs ONNX Conv operation.
///
/// \param node The ONNX node object representing this operation.
///
/// \return The vector containing Ngraph nodes producing output of ONNX convolution
/// operation.
NodeVector conv(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/div.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector div(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Divide>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector div(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Divide>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/elu.cpp
View file @ 6ca1a511
......@@ -38,26 +38,33 @@ namespace ngraph
{
namespace op
{
NodeVector elu(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1);
NodeVector elu(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1);
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
return {std::make_shared<ngraph::op::Maximum>(data, zero_node) +
alpha_node * std::make_shared<ngraph::op::Exp>(
std::make_shared<ngraph::op::Minimum>(data, zero_node)) -
alpha_node};
}
return {std::make_shared<ngraph::op::Maximum>(data, zero_node) +
alpha_node *
std::make_shared<ngraph::op::Exp>(
std::make_shared<ngraph::op::Minimum>(data, zero_node)) -
alpha_node};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/elu.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector elu(const Node& node);
} // namespace op
namespace set_1
{
NodeVector elu(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/equal.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector equal(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Equal>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector equal(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Equal>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/exp.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector exp(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Exp>(node.get_ng_inputs().at(0))};
}
inline NodeVector exp(const Node& node)
{
return {std::make_shared<ngraph::op::Exp>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/flatten.cpp
View file @ 6ca1a511
......@@ -25,19 +25,23 @@ namespace ngraph
{
namespace op
{
NodeVector flatten(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto axis = node.get_attribute_value<int64_t>("axis", 1);
NodeVector flatten(const Node& node)
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto axis = node.get_attribute_value<int64_t>("axis", 1);
ASSERT_VALID_ARGUMENT(node, (axis >= 0) && (axis <= data->get_shape().size()))
<< "provided 'axis' attribute is not valid.";
ASSERT_VALID_ARGUMENT(node, (axis >= 0) && (axis <= data->get_shape().size()))
<< "provided 'axis' attribute is not valid.";
return {reshape::flatten(data, axis)};
}
return {reshape::flatten(data, axis)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/flatten.hpp
View file @ 6ca1a511
......@@ -28,8 +28,13 @@ namespace ngraph
{
namespace op
{
NodeVector flatten(const Node& node);
} // namespace op
namespace set_1
{
NodeVector flatten(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/floor.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector floor(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Floor>(node.get_ng_inputs().at(0))};
}
inline NodeVector floor(const Node& node)
{
return {std::make_shared<ngraph::op::Floor>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/gemm.cpp
View file @ 6ca1a511
......@@ -32,50 +32,58 @@ namespace ngraph
{
namespace op
{
NodeVector gemm(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto input_a = inputs.at(0);
auto input_b = inputs.at(1);
auto input_c = inputs.at(2);
double alpha = node.get_attribute_value<double>("alpha", 1);
double beta = node.get_attribute_value<double>("beta", 1);
auto trans_a = node.get_attribute_value<int64_t>("transA", 0);
auto trans_b = node.get_attribute_value<int64_t>("transB", 0);
if (trans_a != 0)
{
input_a = reshape::transpose(input_a);
}
if (trans_b != 0)
NodeVector gemm(const Node& node)
{
input_b = reshape::transpose(input_b);
NodeVector inputs{node.get_ng_inputs()};
auto input_a = inputs.at(0);
auto input_b = inputs.at(1);
auto input_c = inputs.at(2);
double alpha = node.get_attribute_value<double>("alpha", 1);
double beta = node.get_attribute_value<double>("beta", 1);
auto trans_a = node.get_attribute_value<int64_t>("transA", 0);
auto trans_b = node.get_attribute_value<int64_t>("transB", 0);
if (trans_a != 0)
{
input_a = reshape::transpose(input_a);
}
if (trans_b != 0)
{
input_b = reshape::transpose(input_b);
}
// code from python not implemented in c++ yet.
// reshape_for_matmul(node, input_a, input_b);
std::shared_ptr<ngraph::Node> a_dot_b =
std::make_shared<ngraph::op::Dot>(input_a, input_b);
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(a_dot_b->get_element_type(),
ngraph::Shape{},
std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, a_dot_b->get_shape());
a_dot_b = std::make_shared<ngraph::op::Multiply>(alpha_node, a_dot_b);
std::shared_ptr<ngraph::Node> beta_node =
std::make_shared<ngraph::op::Constant>(input_c->get_element_type(),
ngraph::Shape{},
std::vector<double>{beta});
beta_node = make_broadcast_node(beta_node, input_c->get_shape());
input_c = std::make_shared<ngraph::op::Multiply>(beta_node, input_c);
input_c = make_broadcast_node(input_c, a_dot_b->get_shape());
return {std::make_shared<ngraph::op::Add>(a_dot_b, input_c)};
}
// code from python not implemented in c++ yet.
// reshape_for_matmul(node, input_a, input_b);
std::shared_ptr<ngraph::Node> a_dot_b =
std::make_shared<ngraph::op::Dot>(input_a, input_b);
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
a_dot_b->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, a_dot_b->get_shape());
a_dot_b = std::make_shared<ngraph::op::Multiply>(alpha_node, a_dot_b);
std::shared_ptr<ngraph::Node> beta_node = std::make_shared<ngraph::op::Constant>(
input_c->get_element_type(), ngraph::Shape{}, std::vector<double>{beta});
beta_node = make_broadcast_node(beta_node, input_c->get_shape());
input_c = std::make_shared<ngraph::op::Multiply>(beta_node, input_c);
input_c = make_broadcast_node(input_c, a_dot_b->get_shape());
return {std::make_shared<ngraph::op::Add>(a_dot_b, input_c)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/gemm.hpp
View file @ 6ca1a511
......@@ -28,8 +28,13 @@ namespace ngraph
{
namespace op
{
NodeVector gemm(const Node& node);
} // namespace op
namespace set_1
{
NodeVector gemm(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/greater.hpp
View file @ 6ca1a511
......@@ -28,14 +28,19 @@ namespace ngraph
{
namespace op
{
inline NodeVector greater(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Greater>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector greater(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {
std::make_shared<ngraph::op::Greater>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/hard_sigmoid.cpp
View file @ 6ca1a511
......@@ -35,36 +35,43 @@ namespace ngraph
{
namespace op
{
NodeVector hard_sigmoid(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
NodeVector hard_sigmoid(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 0.2);
double beta = node.get_attribute_value<double>("beta", 0.5);
double alpha = node.get_attribute_value<double>("alpha", 0.2);
double beta = node.get_attribute_value<double>("beta", 0.5);
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> beta_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{beta});
beta_node = make_broadcast_node(beta_node, data->get_shape());
std::shared_ptr<ngraph::Node> beta_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{beta});
beta_node = make_broadcast_node(beta_node, data->get_shape());
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
return {std::make_shared<ngraph::op::Maximum>(
zero_node,
std::make_shared<ngraph::op::Minimum>(one_node,
alpha_node * data + beta_node))};
}
return {std::make_shared<ngraph::op::Maximum>(
zero_node,
std::make_shared<ngraph::op::Minimum>(one_node,
alpha_node * data + beta_node))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/hard_sigmoid.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector hard_sigmoid(const Node& node);
} // namespace op
namespace set_1
{
NodeVector hard_sigmoid(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/identity.hpp
View file @ 6ca1a511
......@@ -26,8 +26,15 @@ namespace ngraph
{
namespace op
{
inline NodeVector identity(const Node& node) { return {node.get_ng_inputs().at(0)}; }
} // namespace op
namespace set_1
{
inline NodeVector identity(const Node& node)
{
return {node.get_ng_inputs().at(0)};
}
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/leaky_relu.cpp
View file @ 6ca1a511
......@@ -38,21 +38,26 @@ namespace ngraph
{
namespace op
{
NodeVector leaky_relu(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 0.01);
NodeVector leaky_relu(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 0.01);
ASSERT_VALID_ARGUMENT(node, ((alpha >= 0) && (alpha <= 1)))
<< " alpha value should be in range (0,1)";
ASSERT_VALID_ARGUMENT(node, ((alpha >= 0) && (alpha <= 1)))
<< " alpha value should be in range (0,1)";
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
return {std::make_shared<ngraph::op::Maximum>(data * alpha_node, data)};
}
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
return {std::make_shared<ngraph::op::Maximum>(data * alpha_node, data)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/leaky_relu.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector leaky_relu(const Node& node);
} // namespace op
namespace set_1
{
NodeVector leaky_relu(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/less.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector less(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Less>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector less(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Less>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/log.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector log(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Log>(node.get_ng_inputs().at(0))};
}
inline NodeVector log(const Node& node)
{
return {std::make_shared<ngraph::op::Log>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/log_softmax.hpp
View file @ 6ca1a511
......@@ -31,13 +31,17 @@ namespace ngraph
{
namespace op
{
inline NodeVector log_softmax(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Log>(softmax(node).at(0))};
}
inline NodeVector log_softmax(const Node& node)
{
return {std::make_shared<ngraph::op::Log>(softmax(node).at(0))};
}
} // namespace op
} // namespace set_1
} // namespace onnx_import
} //namespace op
} // namespace ngraph
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/lrn.cpp
View file @ 6ca1a511
......@@ -27,18 +27,22 @@ namespace ngraph
{
namespace op
{
NodeVector lrn(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1e-4);
double beta = node.get_attribute_value<double>("beta", 0.75);
double bias = node.get_attribute_value<double>("bias", 1);
size_t size = node.get_attribute_value<size_t>("size");
NodeVector lrn(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1e-4);
double beta = node.get_attribute_value<double>("beta", 0.75);
double bias = node.get_attribute_value<double>("bias", 1);
size_t size = node.get_attribute_value<size_t>("size");
return {std::make_shared<ngraph::op::LRN>(data, alpha, beta, bias, size)};
}
return {std::make_shared<ngraph::op::LRN>(data, alpha, beta, bias, size)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/lrn.hpp
View file @ 6ca1a511
......@@ -26,8 +26,12 @@ namespace ngraph
{
namespace op
{
NodeVector lrn(const Node& node);
} // namespace op
namespace set_1
{
NodeVector lrn(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/matmul.hpp
View file @ 6ca1a511
......@@ -27,13 +27,17 @@ namespace ngraph
{
namespace op
{
inline NodeVector matmul(const Node& node)
namespace set_1
{
NodeVector ng_inputs{node.get_ng_inputs()};
return {std::make_shared<ngraph::op::Dot>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector matmul(const Node& node)
{
NodeVector ng_inputs{node.get_ng_inputs()};
return {std::make_shared<ngraph::op::Dot>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/max.hpp
View file @ 6ca1a511
......@@ -28,12 +28,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector max(const Node& node)
namespace set_1
{
return variadic::make_ng_variadic_op<ngraph::op::Maximum>(node);
}
inline NodeVector max(const Node& node)
{
return variadic::make_ng_variadic_op<ngraph::op::Maximum>(node);
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/max_pool.cpp
View file @ 6ca1a511
......@@ -26,12 +26,16 @@ namespace ngraph
{
namespace op
{
NodeVector max_pool(const Node& node)
namespace set_1
{
return convpool::make_ng_pool<ngraph::op::MaxPool>(node);
}
NodeVector max_pool(const Node& node)
{
return convpool::make_ng_pool<ngraph::op::MaxPool>(node);
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/max_pool.hpp
View file @ 6ca1a511
......@@ -26,17 +26,21 @@ namespace ngraph
{
namespace op
{
/**
* @brief Convert ONNX MaxPool operation to an nGraph node.
*
* @param node The ONNX node object representing this operation.
*
* @return The vector containing Ngraph nodes producing output of ONNX MaxPool
* operation.
*/
NodeVector max_pool(const Node& node);
} // namespace op
namespace set_1
{
/**
* @brief Convert ONNX MaxPool operation to an nGraph node.
*
* @param node The ONNX node object representing this operation.
*
* @return The vector containing Ngraph nodes producing output of ONNX MaxPool
* operation.
*/
NodeVector max_pool(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/mean.cpp
View file @ 6ca1a511
......@@ -27,21 +27,25 @@ namespace ngraph
{
namespace op
{
NodeVector mean(const Node& node)
namespace set_1
{
auto sum = variadic::make_ng_variadic_op<ngraph::op::Add>(node).front();
auto shape = sum->get_shape();
NodeVector mean(const Node& node)
{
auto sum = variadic::make_ng_variadic_op<ngraph::op::Add>(node).front();
auto shape = sum->get_shape();
// Create a Constant representing the number of inputs with the same shape as sum
auto count = ngraph::op::Constant::create(
sum->get_element_type(),
shape,
std::vector<int>(shape_size(shape), node.get_ng_inputs().size()));
// Create a Constant representing the number of inputs with the same shape as sum
auto count = ngraph::op::Constant::create(
sum->get_element_type(),
shape,
std::vector<int>(shape_size(shape), node.get_ng_inputs().size()));
return {sum / count};
}
return {sum / count};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/mean.hpp
View file @ 6ca1a511
......@@ -26,9 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector mean(const Node& node);
namespace set_1
{
NodeVector mean(const Node& node);
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/min.hpp
View file @ 6ca1a511
......@@ -28,12 +28,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector min(const Node& node)
namespace set_1
{
return variadic::make_ng_variadic_op<ngraph::op::Minimum>(node);
}
inline NodeVector min(const Node& node)
{
return variadic::make_ng_variadic_op<ngraph::op::Minimum>(node);
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/mul.hpp
View file @ 6ca1a511
......@@ -28,14 +28,19 @@ namespace ngraph
{
namespace op
{
inline NodeVector mul(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Multiply>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector mul(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {
std::make_shared<ngraph::op::Multiply>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/neg.hpp
View file @ 6ca1a511
......@@ -27,8 +27,12 @@ namespace ngraph
{
namespace op
{
inline NodeVector neg(const Node& node) { return {-node.get_ng_inputs().at(0)}; }
} // namespace op
namespace set_1
{
inline NodeVector neg(const Node& node) { return {-node.get_ng_inputs().at(0)}; }
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/not.hpp
View file @ 6ca1a511
......@@ -28,12 +28,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector logical_not(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Not>(node.get_ng_inputs().at(0))};
}
inline NodeVector logical_not(const Node& node)
{
return {std::make_shared<ngraph::op::Not>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/or.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector logical_or(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Or>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector logical_or(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Or>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/pow.hpp
View file @ 6ca1a511
......@@ -28,14 +28,18 @@ namespace ngraph
{
namespace op
{
inline NodeVector pow(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Power>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector pow(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Power>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/prelu.cpp
View file @ 6ca1a511
......@@ -37,31 +37,35 @@ namespace ngraph
{
namespace op
{
NodeVector prelu(const Node& node)
namespace set_1
{
NodeVector ng_inputs{node.get_ng_inputs()};
auto data = ng_inputs.at(0);
auto data_shape = data->get_shape();
std::shared_ptr<ngraph::Node> slope = ng_inputs.at(1);
auto slope_shape = slope->get_shape();
if ((slope_shape.size() == 1) && (slope_shape.at(0) != 1))
{
auto it =
std::find(std::begin(data_shape), std::end(data_shape), slope_shape.at(0));
auto index = std::distance(std::begin(data_shape), it);
slope = make_broadcast_node(slope, data->get_shape(), index);
}
else
NodeVector prelu(const Node& node)
{
auto params = numpy_style_broadcast_for_binary_operation(slope, data);
slope = params.at(0);
NodeVector ng_inputs{node.get_ng_inputs()};
auto data = ng_inputs.at(0);
auto data_shape = data->get_shape();
std::shared_ptr<ngraph::Node> slope = ng_inputs.at(1);
auto slope_shape = slope->get_shape();
if ((slope_shape.size() == 1) && (slope_shape.at(0) != 1))
{
auto it = std::find(
std::begin(data_shape), std::end(data_shape), slope_shape.at(0));
auto index = std::distance(std::begin(data_shape), it);
slope = make_broadcast_node(slope, data->get_shape(), index);
}
else
{
auto params = numpy_style_broadcast_for_binary_operation(slope, data);
slope = params.at(0);
}
return {std::make_shared<ngraph::op::Maximum>(data * slope, data)};
}
return {std::make_shared<ngraph::op::Maximum>(data * slope, data)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/prelu.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector prelu(const Node& node);
} // namespace op
namespace set_1
{
NodeVector prelu(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/reciprocal.cpp
View file @ 6ca1a511
......@@ -31,18 +31,22 @@ namespace ngraph
{
namespace op
{
NodeVector reciprocal(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
NodeVector reciprocal(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
return {one_node / data};
}
return {one_node / data};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/reciprocal.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector reciprocal(const Node& node);
} // namespace op
namespace set_1
{
NodeVector reciprocal(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/reduce.cpp
View file @ 6ca1a511
......@@ -32,29 +32,35 @@ namespace ngraph
{
namespace op
{
NodeVector reduce_mean(const Node& node)
namespace set_1
{
auto input_shape = node.get_ng_inputs().at(0)->get_shape();
auto reduction_axes = reduction::detail::get_reduction_axes(node);
std::size_t elem_count_product =
std::accumulate(std::begin(reduction_axes),
std::end(reduction_axes),
1UL,
[&input_shape](const std::size_t& a, const std::size_t& b) {
return a * input_shape.at(b);
});
auto sum_node = reduction::make_ng_reduction_op<ngraph::op::Sum>(
node, node.get_ng_inputs().at(0));
auto const_node = std::make_shared<ngraph::op::Constant>(
sum_node->get_element_type(),
Shape{},
std::vector<std::size_t>{elem_count_product});
auto broadcasted_const_node =
make_broadcast_node(const_node, sum_node->get_shape());
return {std::make_shared<ngraph::op::Divide>(sum_node, broadcasted_const_node)};
}
} // namespace op
} // namespace onnx_import
NodeVector reduce_mean(const Node& node)
{
auto input_shape = node.get_ng_inputs().at(0)->get_shape();
auto reduction_axes = reduction::detail::get_reduction_axes(node);
std::size_t elem_count_product =
std::accumulate(std::begin(reduction_axes),
std::end(reduction_axes),
1UL,
[&input_shape](const std::size_t& a, const std::size_t& b) {
return a * input_shape.at(b);
});
auto sum_node = reduction::make_ng_reduction_op<ngraph::op::Sum>(
node, node.get_ng_inputs().at(0));
auto const_node = std::make_shared<ngraph::op::Constant>(
sum_node->get_element_type(),
Shape{},
std::vector<std::size_t>{elem_count_product});
auto broadcasted_const_node =
make_broadcast_node(const_node, sum_node->get_shape());
return {std::make_shared<ngraph::op::Divide>(sum_node, broadcasted_const_node)};
}
} // namespace set_1
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/relu.hpp
View file @ 6ca1a511
......@@ -27,13 +27,17 @@ namespace ngraph
{
namespace op
{
inline NodeVector relu(const Node& node)
namespace set_1
{
NodeVector ng_inputs{node.get_ng_inputs()};
return {std::make_shared<ngraph::op::Relu>(ng_inputs.at(0))};
}
inline NodeVector relu(const Node& node)
{
NodeVector ng_inputs{node.get_ng_inputs()};
return {std::make_shared<ngraph::op::Relu>(ng_inputs.at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/reshape.cpp
View file @ 6ca1a511
......@@ -33,39 +33,45 @@ namespace ngraph
{
namespace op
{
NodeVector reshape(const Node& node)
namespace set_1
{
NodeVector ng_inputs{node.get_ng_inputs()};
auto data = ng_inputs.at(0);
auto data_shape = data->get_shape();
NodeVector reshape(const Node& node)
{
NodeVector ng_inputs{node.get_ng_inputs()};
auto data = ng_inputs.at(0);
auto data_shape = data->get_shape();
auto output_shape = node.get_attribute_value<std::vector<std::size_t>>("shape", {});
auto output_shape =
node.get_attribute_value<std::vector<std::size_t>>("shape", {});
// If no shape argument (opset >= 5) and there is second input.
if (output_shape.empty() && ng_inputs.size() == 2)
{
// Currently only support Constant node.
ASSERT_IS_SUPPORTED(node, ng_inputs.at(1)->description() == "Constant")
<< "doesn't support shape input of other type than Constant.";
// If no shape argument (opset >= 5) and there is second input.
if (output_shape.empty() && ng_inputs.size() == 2)
{
// Currently only support Constant node.
ASSERT_IS_SUPPORTED(node, ng_inputs.at(1)->description() == "Constant")
<< "doesn't support shape input of other type than Constant.";
auto output_shape_node =
std::dynamic_pointer_cast<ngraph::op::Constant>(ng_inputs.at(1));
output_shape = output_shape_node->get_vector<std::size_t>();
}
// Do nothing if there is no shape argument nor second node input.
else if (output_shape.empty())
{
return {data};
auto output_shape_node =
std::dynamic_pointer_cast<ngraph::op::Constant>(ng_inputs.at(1));
output_shape = output_shape_node->get_vector<std::size_t>();
}
// Do nothing if there is no shape argument nor second node input.
else if (output_shape.empty())
{
return {data};
}
output_shape =
reshape::infer_dimensions(node.get_name(), data_shape, output_shape);
return {std::make_shared<ngraph::op::Reshape>(
data,
reshape::get_default_axis_vector(data_shape.size()),
Shape{output_shape})};
}
output_shape = reshape::infer_dimensions(node.get_name(), data_shape, output_shape);
return {std::make_shared<ngraph::op::Reshape>(
data,
reshape::get_default_axis_vector(data_shape.size()),
Shape{output_shape})};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/reshape.hpp
View file @ 6ca1a511
......@@ -26,16 +26,20 @@ namespace ngraph
{
namespace op
{
///
/// \brief Reshape the input tensor similar to numpy.reshape.
///
/// \param[in] node The ONNX node representing this operation.
///
/// \return Ngraph node representing this operation.
///
NodeVector reshape(const Node& node);
} // namespace op
namespace set_1
{
///
/// \brief Reshape the input tensor similar to numpy.reshape.
///
/// \param[in] node The ONNX node representing this operation.
///
/// \return Ngraph node representing this operation.
///
NodeVector reshape(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/selu.cpp
View file @ 6ca1a511
......@@ -41,32 +41,41 @@ namespace ngraph
{
namespace op
{
NodeVector selu(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1.67326319217681884765625);
double gamma = node.get_attribute_value<double>("gamma", 1.05070102214813232421875);
NodeVector selu(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha =
node.get_attribute_value<double>("alpha", 1.67326319217681884765625);
double gamma =
node.get_attribute_value<double>("gamma", 1.05070102214813232421875);
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> gamma_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{gamma});
gamma_node = make_broadcast_node(gamma_node, data->get_shape());
std::shared_ptr<ngraph::Node> gamma_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{gamma});
gamma_node = make_broadcast_node(gamma_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
std::shared_ptr<ngraph::Node> zero_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{0});
zero_node = make_broadcast_node(zero_node, data->get_shape());
return {gamma_node *
(std::make_shared<ngraph::op::Maximum>(data, zero_node) +
alpha_node * std::make_shared<ngraph::op::Exp>(
std::make_shared<ngraph::op::Minimum>(data, zero_node)) -
alpha_node)};
}
return {gamma_node * (std::make_shared<ngraph::op::Maximum>(data, zero_node) +
alpha_node * std::make_shared<ngraph::op::Exp>(
std::make_shared<ngraph::op::Minimum>(
data, zero_node)) -
alpha_node)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/selu.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector selu(const Node& node);
} // namespace op
namespace set_1
{
NodeVector selu(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/shape.cpp
View file @ 6ca1a511
......@@ -30,16 +30,20 @@ namespace ngraph
{
namespace op
{
NodeVector shape(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
auto data_shape = data->get_shape();
NodeVector shape(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
auto data_shape = data->get_shape();
return {std::make_shared<ngraph::op::Constant>(
ngraph::element::i64, Shape{data_shape.size()}, data_shape)};
}
return {std::make_shared<ngraph::op::Constant>(
ngraph::element::i64, Shape{data_shape.size()}, data_shape)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/shape.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector shape(const Node& node);
} // namespace op
namespace set_1
{
NodeVector shape(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/sigmoid.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector sigmoid(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Sigmoid>(node.get_ng_inputs().at(0))};
}
inline NodeVector sigmoid(const Node& node)
{
return {std::make_shared<ngraph::op::Sigmoid>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/slice.cpp
View file @ 6ca1a511
......@@ -35,32 +35,37 @@ namespace ngraph
{
namespace op
{
NodeVector slice(const Node& node)
namespace set_1
{
std::shared_ptr<ngraph::Node> data = node.get_ng_inputs().at(0);
Shape data_shape = data->get_shape();
NodeVector slice(const Node& node)
{
std::shared_ptr<ngraph::Node> data = node.get_ng_inputs().at(0);
Shape data_shape = data->get_shape();
auto starts = node.get_attribute_value<std::vector<int64_t>>("starts");
auto ends = node.get_attribute_value<std::vector<int64_t>>("ends");
auto starts = node.get_attribute_value<std::vector<int64_t>>("starts");
auto ends = node.get_attribute_value<std::vector<int64_t>>("ends");
auto axes = node.get_attribute_value<std::vector<int64_t>>(
"axes", common::get_monotonic_range<int64_t>(data_shape.size()));
auto axes = node.get_attribute_value<std::vector<int64_t>>(
"axes", common::get_monotonic_range<int64_t>(data_shape.size()));
Shape lower_bounds(data_shape.size());
Shape upper_bounds = data_shape;
Shape lower_bounds(data_shape.size());
Shape upper_bounds = data_shape;
for (auto idx = 0; idx < axes.size(); ++idx)
{
size_t axis = axes.at(idx);
lower_bounds.at(axis) =
get_valid_array_idx(starts.at(idx), data_shape.at(axis));
upper_bounds.at(axis) = get_valid_array_idx(ends.at(idx), data_shape.at(axis));
for (auto idx = 0; idx < axes.size(); ++idx)
{
size_t axis = axes.at(idx);
lower_bounds.at(axis) =
get_valid_array_idx(starts.at(idx), data_shape.at(axis));
upper_bounds.at(axis) =
get_valid_array_idx(ends.at(idx), data_shape.at(axis));
}
return {std::make_shared<ngraph::op::Slice>(data, lower_bounds, upper_bounds)};
}
return {std::make_shared<ngraph::op::Slice>(data, lower_bounds, upper_bounds)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/slice.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector slice(const Node& node);
} // namespace op
namespace set_1
{
NodeVector slice(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/softmax.cpp
View file @ 6ca1a511
......@@ -27,31 +27,35 @@ namespace ngraph
{
namespace op
{
NodeVector softmax(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
NodeVector softmax(const Node& node)
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
int axis = node.get_attribute_value<int64_t>("axis", 1);
int axis = node.get_attribute_value<int64_t>("axis", 1);
if (axis < 0)
{
axis = data_shape.size() + axis;
}
if (axis < 0)
{
axis = data_shape.size() + axis;
}
ASSERT_VALID_ARGUMENT(node, axis < data_shape.size())
<< "provided 'axis' value:" << axis
<< " is out of input tensor dimensions range.";
ASSERT_VALID_ARGUMENT(node, axis < data_shape.size())
<< "provided 'axis' value:" << axis
<< " is out of input tensor dimensions range.";
// create vector of capacity data_dimensions - axis_divider position
std::vector<size_t> axes(data_shape.size() - axis);
std::iota(std::begin(axes), std::end(axes), axis);
return {std::make_shared<ngraph::op::Softmax>(data, axes)};
}
// create vector of capacity data_dimensions - axis_divider position
std::vector<size_t> axes(data_shape.size() - axis);
std::iota(std::begin(axes), std::end(axes), axis);
return {std::make_shared<ngraph::op::Softmax>(data, axes)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
} // namespace onnx_import
} // namespace ngraph
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/softmax.hpp
View file @ 6ca1a511
......@@ -26,9 +26,14 @@ namespace ngraph
{
namespace op
{
NodeVector softmax(const Node& node);
} // namespace op
namespace set_1
{
NodeVector softmax(const Node& node);
} // namespace onnx_import
} // namespace set_1
} // namespace ngraph
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/softplus.cpp
View file @ 6ca1a511
......@@ -33,19 +33,23 @@ namespace ngraph
{
namespace op
{
NodeVector softplus(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
NodeVector softplus(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
return {std::make_shared<ngraph::op::Log>(std::make_shared<ngraph::op::Exp>(data) +
one_node)};
}
return {std::make_shared<ngraph::op::Log>(
std::make_shared<ngraph::op::Exp>(data) + one_node)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/softplus.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector softplus(const Node& node);
} // namespace op
namespace set_1
{
NodeVector softplus(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/softsign.cpp
View file @ 6ca1a511
......@@ -33,18 +33,22 @@ namespace ngraph
{
namespace op
{
NodeVector softsign(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
NodeVector softsign(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
std::shared_ptr<ngraph::Node> one_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), Shape{}, std::vector<double>{1});
one_node = make_broadcast_node(one_node, data->get_shape());
return {data / (std::make_shared<ngraph::op::Abs>(data) + one_node)};
}
return {data / (std::make_shared<ngraph::op::Abs>(data) + one_node)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/softsign.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector softsign(const Node& node);
} // namespace op
namespace set_1
{
NodeVector softsign(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/split.cpp
View file @ 6ca1a511
......@@ -80,79 +80,84 @@ namespace ngraph
namespace op
{
namespace detail
namespace set_1
{
template <typename T>
inline T get_valid_array_index(T left, T right)
namespace detail
{
return (left >= 0) ? std::min(left, right)
: std::max(static_cast<T>(0), right + left);
}
template <typename T>
inline T get_valid_array_index(T left, T right)
{
return (left >= 0) ? std::min(left, right)
: std::max(static_cast<T>(0), right + left);
}
inline std::shared_ptr<ngraph::op::Slice>
make_ng_slice(const std::shared_ptr<ngraph::Node>& node,
std::vector<std::size_t> axes,
std::vector<std::size_t> starts,
std::vector<std::size_t> ends)
{
std::vector<std::size_t> upper_bounds{node->get_shape()};
std::vector<std::size_t> lower_bounds(upper_bounds.size());
for (std::size_t index{0}; index < axes.size(); ++index)
inline std::shared_ptr<ngraph::op::Slice>
make_ng_slice(const std::shared_ptr<ngraph::Node>& node,
std::vector<std::size_t> axes,
std::vector<std::size_t> starts,
std::vector<std::size_t> ends)
{
std::size_t axis{axes.at(index)};
lower_bounds.at(axis) =
get_valid_array_index(starts.at(index), node->get_shape().at(axis));
upper_bounds.at(axis) =
get_valid_array_index(ends.at(index), node->get_shape().at(axis));
std::vector<std::size_t> upper_bounds{node->get_shape()};
std::vector<std::size_t> lower_bounds(upper_bounds.size());
for (std::size_t index{0}; index < axes.size(); ++index)
{
std::size_t axis{axes.at(index)};
lower_bounds.at(axis) =
get_valid_array_index(starts.at(index), node->get_shape().at(axis));
upper_bounds.at(axis) =
get_valid_array_index(ends.at(index), node->get_shape().at(axis));
}
return std::make_shared<ngraph::op::Slice>(
node, lower_bounds, upper_bounds);
}
return std::make_shared<ngraph::op::Slice>(node, lower_bounds, upper_bounds);
}
} // namespace detail
} // namespace detail
NodeVector split(const Node& node)
{
std::shared_ptr<ngraph::Node> input = node.get_ng_inputs().at(0);
std::size_t count_outputs{node.get_output_names().size()};
int64_t axis{node.get_attribute_value<int64_t>("axis", 0)};
std::size_t axis_to_split{static_cast<std::size_t>(axis)};
if (axis < 0)
{
axis_to_split = input->get_shape().size() + axis;
}
else if (axis_to_split >= input->get_shape().size())
{
throw error::op::split::OutOfRange{node.get_name()};
}
std::size_t length_axis_to_split{input->get_shape().at(axis_to_split)};
std::vector<std::size_t> length_parts;
try
{
length_parts = node.get_attribute_value<std::vector<std::size_t>>("split");
}
catch (const std::exception&)
NodeVector split(const Node& node)
{
if (length_axis_to_split % count_outputs)
std::shared_ptr<ngraph::Node> input = node.get_ng_inputs().at(0);
std::size_t count_outputs{node.get_output_names().size()};
int64_t axis{node.get_attribute_value<int64_t>("axis", 0)};
std::size_t axis_to_split{static_cast<std::size_t>(axis)};
if (axis < 0)
{
throw error::op::split::Parts{
node.get_name(), count_outputs, length_axis_to_split};
axis_to_split = input->get_shape().size() + axis;
}
else if (axis_to_split >= input->get_shape().size())
{
throw error::op::split::OutOfRange{node.get_name()};
}
std::size_t length_axis_to_split{input->get_shape().at(axis_to_split)};
std::vector<std::size_t> length_parts;
try
{
length_parts = node.get_attribute_value<std::vector<std::size_t>>("split");
}
catch (const std::exception&)
{
if (length_axis_to_split % count_outputs)
{
throw error::op::split::Parts{
node.get_name(), count_outputs, length_axis_to_split};
}
length_parts.assign(count_outputs, length_axis_to_split / count_outputs);
}
length_parts.assign(count_outputs, length_axis_to_split / count_outputs);
}
std::size_t start_index{0};
NodeVector outputs;
for (const auto& length_part : length_parts)
{
std::size_t end_index{start_index + length_part};
outputs.push_back(
detail::make_ng_slice(input, {axis_to_split}, {start_index}, {end_index}));
start_index = end_index;
std::size_t start_index{0};
NodeVector outputs;
for (const auto& length_part : length_parts)
{
std::size_t end_index{start_index + length_part};
outputs.push_back(detail::make_ng_slice(
input, {axis_to_split}, {start_index}, {end_index}));
start_index = end_index;
}
return outputs;
}
return outputs;
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/split.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector split(const Node& node);
} // namespace op
namespace set_1
{
NodeVector split(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/sqrt.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector sqrt(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Sqrt>(node.get_ng_inputs().at(0))};
}
inline NodeVector sqrt(const Node& node)
{
return {std::make_shared<ngraph::op::Sqrt>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/squeeze.cpp
View file @ 6ca1a511
......@@ -30,38 +30,42 @@ namespace ngraph
{
namespace op
{
NodeVector squeeze(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
auto axes = node.get_attribute_value<std::vector<uint64_t>>("axes", {});
if (axes.empty())
NodeVector squeeze(const Node& node)
{
for (auto index = 0; index < data_shape.size(); ++index)
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
auto axes = node.get_attribute_value<std::vector<uint64_t>>("axes", {});
if (axes.empty())
{
if (data_shape.at(index) == 1)
for (auto index = 0; index < data_shape.size(); ++index)
{
axes.push_back(index);
if (data_shape.at(index) == 1)
{
axes.push_back(index);
}
}
}
}
std::sort(std::begin(axes), std::end(axes), std::greater<uint64_t>());
std::sort(std::begin(axes), std::end(axes), std::greater<uint64_t>());
AxisVector input_order{reshape::get_default_axis_vector(data_shape.size())};
AxisVector input_order{reshape::get_default_axis_vector(data_shape.size())};
for (auto axis : axes)
{
data_shape.erase(std::next(std::begin(data_shape), axis));
for (auto axis : axes)
{
data_shape.erase(std::next(std::begin(data_shape), axis));
}
return {std::make_shared<ngraph::op::Reshape>(data, input_order, data_shape)};
}
return {std::make_shared<ngraph::op::Reshape>(data, input_order, data_shape)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
} // namespace onnx_import
} // namespace ngraph
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/squeeze.hpp
View file @ 6ca1a511
......@@ -26,9 +26,14 @@ namespace ngraph
{
namespace op
{
NodeVector squeeze(const Node& node);
} // namespace op
namespace set_1
{
NodeVector squeeze(const Node& node);
} // namespace onnx_import
} // namespace set_1
} // namespace ngraph
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/sub.hpp
View file @ 6ca1a511
......@@ -28,14 +28,19 @@ namespace ngraph
{
namespace op
{
inline NodeVector sub(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {std::make_shared<ngraph::op::Subtract>(ng_inputs.at(0), ng_inputs.at(1))};
}
inline NodeVector sub(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
return {
std::make_shared<ngraph::op::Subtract>(ng_inputs.at(0), ng_inputs.at(1))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/sum.hpp
View file @ 6ca1a511
......@@ -28,12 +28,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector sum(const Node& node)
namespace set_1
{
return variadic::make_ng_variadic_op<ngraph::op::Add>(node);
}
inline NodeVector sum(const Node& node)
{
return variadic::make_ng_variadic_op<ngraph::op::Add>(node);
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/tanh.hpp
View file @ 6ca1a511
......@@ -29,12 +29,16 @@ namespace ngraph
{
namespace op
{
inline NodeVector tanh(const Node& node)
namespace set_1
{
return {std::make_shared<ngraph::op::Tanh>(node.get_ng_inputs().at(0))};
}
inline NodeVector tanh(const Node& node)
{
return {std::make_shared<ngraph::op::Tanh>(node.get_ng_inputs().at(0))};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/thresholded_relu.cpp
View file @ 6ca1a511
......@@ -36,22 +36,27 @@ namespace ngraph
{
namespace op
{
NodeVector thresholded_relu(const Node& node)
namespace set_1
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1.0);
NodeVector thresholded_relu(const Node& node)
{
auto data = node.get_ng_inputs().at(0);
double alpha = node.get_attribute_value<double>("alpha", 1.0);
std::shared_ptr<ngraph::Node> alpha_node = std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
std::shared_ptr<ngraph::Node> alpha_node =
std::make_shared<ngraph::op::Constant>(
data->get_element_type(), ngraph::Shape{}, std::vector<double>{alpha});
alpha_node = make_broadcast_node(alpha_node, data->get_shape());
auto data_map = std::make_shared<ngraph::op::Convert>(
std::make_shared<ngraph::op::Greater>(data, alpha_node),
data->get_element_type());
return {data * data_map};
}
auto data_map = std::make_shared<ngraph::op::Convert>(
std::make_shared<ngraph::op::Greater>(data, alpha_node),
data->get_element_type());
return {data * data_map};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/thresholded_relu.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector thresholded_relu(const Node& node);
} // namespace op
namespace set_1
{
NodeVector thresholded_relu(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/transpose.cpp
View file @ 6ca1a511
......@@ -28,17 +28,22 @@ namespace ngraph
{
namespace op
{
NodeVector transpose(const Node& node)
namespace set_1
{
std::shared_ptr<ngraph::Node> data = node.get_ng_inputs().at(0);
NodeVector transpose(const Node& node)
{
std::shared_ptr<ngraph::Node> data = node.get_ng_inputs().at(0);
auto permute_axes = node.get_attribute_value<std::vector<std::size_t>>("perm", {});
auto permute_axes =
node.get_attribute_value<std::vector<std::size_t>>("perm", {});
return {(permute_axes.empty()) ? reshape::transpose(data)
: reshape::reorder_axes(data, permute_axes)};
}
return {(permute_axes.empty()) ? reshape::transpose(data)
: reshape::reorder_axes(data, permute_axes)};
}
} // namespace op
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/transpose.hpp
View file @ 6ca1a511
......@@ -26,8 +26,13 @@ namespace ngraph
{
namespace op
{
NodeVector transpose(const Node& node);
} // namespace op
namespace set_1
{
NodeVector transpose(const Node& node);
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/frontend/onnx_import/op/unsqueeze.cpp
View file @ 6ca1a511
......@@ -28,32 +28,36 @@ namespace ngraph
{
namespace op
{
NodeVector unsqueeze(const Node& node)
namespace set_1
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
auto axes = node.get_attribute_value<std::vector<int64_t>>("axes");
NodeVector unsqueeze(const Node& node)
{
NodeVector inputs{node.get_ng_inputs()};
auto data = inputs.at(0);
auto data_shape = data->get_shape();
auto axes = node.get_attribute_value<std::vector<int64_t>>("axes");
ASSERT_VALID_ARGUMENT(node, !axes.empty()) << "'axes' attribute is mandatory.";
ASSERT_VALID_ARGUMENT(node, !axes.empty()) << "'axes' attribute is mandatory.";
std::sort(std::begin(axes), std::end(axes), std::less<int64_t>());
std::sort(std::begin(axes), std::end(axes), std::less<int64_t>());
AxisVector input_order{reshape::get_default_axis_vector(data_shape.size())};
AxisVector input_order{reshape::get_default_axis_vector(data_shape.size())};
for (auto axis : axes)
{
ASSERT_VALID_ARGUMENT(node, axis >= 0 && axis <= data_shape.size())
<< "provided 'axes' attribute is not valid.";
for (auto axis : axes)
{
ASSERT_VALID_ARGUMENT(node, axis >= 0 && axis <= data_shape.size())
<< "provided 'axes' attribute is not valid.";
data_shape.insert(std::next(std::begin(data_shape), axis), 1);
}
data_shape.insert(std::next(std::begin(data_shape), axis), 1);
return {std::make_shared<ngraph::op::Reshape>(data, input_order, data_shape)};
}
return {std::make_shared<ngraph::op::Reshape>(data, input_order, data_shape)};
}
} // namespace set_1
} // namespace op
} //namespace op
} // namespace onnx_import
} // namespace onnx_import
} // namespace ngraph
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/unsqueeze.hpp
View file @ 6ca1a511
......@@ -26,9 +26,14 @@ namespace ngraph
{
namespace op
{
NodeVector unsqueeze(const Node& node);
} // namespace op
namespace set_1
{
NodeVector unsqueeze(const Node& node);
} // namespace onnx_import
} // namespace set_1
} // namespace ngraph
} //namespace op
} // namespace onnx_import
} // namespace ngraph
src/ngraph/frontend/onnx_import/op/xor.hpp
View file @ 6ca1a511
......@@ -30,20 +30,24 @@ namespace ngraph
{
namespace op
{
inline NodeVector logical_xor(const Node& node)
namespace set_1
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
auto left = ng_inputs.at(0);
auto not_left = std::make_shared<ngraph::op::Not>(left);
auto right = ng_inputs.at(1);
auto not_right = std::make_shared<ngraph::op::Not>(right);
return {std::make_shared<ngraph::op::Or>(
std::make_shared<ngraph::op::And>(left, not_right),
std::make_shared<ngraph::op::And>(not_left, right))};
}
} // namespace op
inline NodeVector logical_xor(const Node& node)
{
NodeVector ng_inputs{
numpy_style_broadcast_for_binary_operation(node.get_ng_inputs())};
auto left = ng_inputs.at(0);
auto not_left = std::make_shared<ngraph::op::Not>(left);
auto right = ng_inputs.at(1);
auto not_right = std::make_shared<ngraph::op::Not>(right);
return {std::make_shared<ngraph::op::Or>(
std::make_shared<ngraph::op::And>(left, not_right),
std::make_shared<ngraph::op::And>(not_left, right))};
}
} // namespace set_1
} //namespace op
} // namespace onnx_import
......
src/ngraph/graph_util.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/graph_util.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/node.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/CMakeLists.txt
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/builder/softmax.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/kernel/result.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/kernel/softmax.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/mkldnn_utils.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/pass/cpu_layout.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/pass/cpu_reshape_sinking.cpp 0 → 100644
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/cpu/pass/cpu_reshape_sinking.hpp 0 → 100644
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/cuda_emitter.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/cuda_emitter.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/cudnn_emitter.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/cudnn_emitter.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/gpu_emitter.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/gpu_emitter.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/gpu/unit_test.manifest
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/runtime/performance_counter.hpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/util.cpp
View file @ 6ca1a511
This diff is collapsed.
src/ngraph/util.hpp
View file @ 6ca1a511
This diff is collapsed.
test/CMakeLists.txt
View file @ 6ca1a511
This diff is collapsed.
test/backend_test.in.cpp
View file @ 6ca1a511
This diff is collapsed.
test/coordinate.cpp
View file @ 6ca1a511
This diff is collapsed.
test/cpu_fusion.cpp
View file @ 6ca1a511
This diff is collapsed.
test/cpu_reshape_sinking.cpp 0 → 100644
View file @ 6ca1a511
This diff is collapsed.
test/cpu_test.cpp
View file @ 6ca1a511
This diff is collapsed.
test/gpu_test.cpp
View file @ 6ca1a511
This diff is collapsed.
test/includes.cpp
View file @ 6ca1a511
This diff is collapsed.
test/models/tf_conv_mnist_nhwc.json 0 → 100644
View file @ 6ca1a511
This diff is collapsed.
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment