* add convolution in progress

* enable 1 test

* convolution in progress

* use filter descripter

* filter discreptor bug fix

* tensor format

* add missed dimension calculator

* forward convolution 4d without dilation and padding working

* data dilation(deconvolution) and enable some test

* add backprop convolution data and filter

* backprop can compile

* pass unit test, but still have problem on padding

* 2d, symmtric padding, no data dilation works now

* clean up code

* extend gpu convolution to nd

* fix some bugs

* working version for upto 3d convolution, code format.

* remove nunecessary changes

* add restriction for data dilation and asymmetric padding

* clang format

* support upto 3D convolution for now

* change comments to not implemented

* change comments to not implemented

* add quary for additional GPU workspace for convolution

* clang format

* code format

* using row_major_strides

* using join

* fix bug for join

* refactor dimension calculation

* Enable BatchNorm op

* Enable function call op

* Enable get output element op

* new backend API in graph partition

* update API

*  zero dimension tensor elimination init

* more ops + refactor + tests

* revert pattern.cpp

* add internal zero-length test

* address Scott's feedback

* fix comp errors

* proper static init

* get rid of unique-ptr

* refactor hashmap into virtual get_default_values on op classes

* fix formatting

* back out api change

* Remove the no longer supported alternative installation method for
python binding.

* Put back CMakeLists.txt as it is used by travis ci dockerfile.

* Remove python/CMakeLists.txt and update Travis CI

* CPU: Optimize 4D "nGraph" Reshapes (shuffle+reshape)

* CPU: Add kernel sources

* CPU: Replace 2D with 3D reshape

* CPU: Fixes

* CPU: Simplify

* remove parameter check from Function::get_ops()

* create validate pass to hold parameter validation

* unskipping passing gpu tests

* skipping failing gpu tests

* import pytest as needed

* fix style issues

* unskip passing test

* add additional skip reason, unable to compile

* repacking recurrent matching as a standalone class

* RecurrentMatcher

* add a getter for root node

* address Scott's feedback

* WIP editing so far for review and feedback

* Add missing env var export for neon install new process

* Add modified venv setup for TF

* More edits for FW integration and landpage

* Revise from PR feedback

* More PR feedback and editing for clarity

* Minor rewording, clearer explanation

* Final pass edit

* more editing

* adding support for GPU backend to contrib/docker

added gpu dockerfiles

renamed Dockerfile for centos74

fixed NGRAPH_GPU_ENABLE cmake flag name

* Check for GPU support on the host system and fall back to CPU if not present

* removed double option for PREBUILT_LLVM

* updated README.md with additional references for GPU support

* added clarifying comments

cleaned up duplicate settings

* removed deprecated targets from the contrib/docker/Makefile

* resolved absolute vs. conditional assignment for variables based on reference OS

* removed example using a custom DOCKERFILE from README file

* force backend compile() to make a copy of the graph

fix copy_with_new_args on ops that have function pointers internal

update unit test for new backend API

add unit test for multiple simulataneous backends

* move get_subdevices virtual method to Manager class

* update GPU to latest

* update call methods

* add remove_compiled_function()

[Py] Change Python version for tox

* CPU: Fuse zero-padded convolution backprop filters

* CPU: Add a testcase for zero-padded convolution backprop filters fusion

* initial support for recurring matching

* fix a bug where patterns weren't populated w/ matched nodes; add recurrent tests

* add a missing newline

* address feedback

* fix function comment

* Add/update Python wrappers for nGraph operations.

- NotEqual, OneHot, Power, Sqrt, Relu, Sign, Sin, Sinh, Tan, Subtract, Select, Tanh, Sum, Reduce,
Softmax, ReplaceSlice, Reverse
- Add UT for Relu, Sign, Sin, Sinh, Sqrt, Tan, Tanh,

* Add UT for cases when Cos and Sin are giving incorrect results.

* Alphabetically sorted imports.

* Small refactoring.

- Update docstrings
- Remove unnecesary auxiliary local variable.

* make Input descriptors node owners

* rename src_node to m_src_node

* Update README to make it consistent with ngraph-neon

* Add instruction to clone pybind in README

* Address feedback on README

* visualization tracing

* visualize -> m_visualize. add a programmatic way to enable visualization. tweak pass names

- Enable Padding op
- Supress multiline comment warning
- improve tox configuration

* enable TensorView to use pre-allocated mem

* proper check for nullptr

* add unittest for custom mem with tensorview and feedback

* minor fix from feedback

* support GPU TensorView custom mem

* feedback fix and code format

* CPU: Optimize Sum reductions

* CPU: Optimize 1D reduce all case

* CPU: Optimize 4D reduce all sum

* CPU: Tweaks

* Formatting fixes

* Enable Travis CI

* Enable Travis CI - update README

* Travis CI - Parallel testing

* saved a simplified contrib/docker/Makefile and helper scripts

* fixed flow for make targets

* restored original definition for setting the PARALLEL option on the command line per github comments

* remove double build for make install targets

* added a save for the ngraph_dist_gcc.tgz to maintain existing behavior

* fixed passing the PARALLEL value to the make targets

* integrated latest working build-ngraph-and-test script

* integrated the latest working Makefile

* removed reference to the THIRD_PARTY_CACHE_DIR (for future)

* updated the contrib/docker/README.md file

* refactor Adjoints to support multi-output ops

* passing tests

* switch to generate_adjoints(deltas) and backprop_node

* remove debugging code

* fix error msg

* fix typo adjoitns

* fix comp errors in mnist_mlp

* [Py]Fix problem with double set layout

* Extend UT for coverage double set layout

* add layouts and users check
* add convrelu handler

* Fix clang warnings on macos

* Conditionalize warning on Apple clang version.