* Added elementwise-add test

* Fix typo

* Fixed elem-wise ops for lists with len>2

* add tutorial for mali gpu

* update submodule version

* fix doc error

* fix server error

* fix server in test

* Explain how to generate module library

* Small fix

* Add conv2d max_pool backward op

* Added tests

* Fix testing

* Address comments

* Change dot to matmul

* Address comments

* Break down indicator function

* Make greater, less numpy compatible

* Keras keeps renaming properties. Update frontend for keras 2.1.3 compatibility

* Add error message when inbound_nodes is not found

* [CI] turn on keras frontend test

* fix

* using tensorflow cpu version

* make shape inference of BatchNorm layout neutral

* refactor to use the axis variable to do BatchNorm shape inference

* refactor to use the axis variable to do BatchNorm shape inference

* add unittest to the axis param for batch norm shape inference

* Add support for keras upsampling

* Fix code formatting

* Fix indentation

* Fix indentation round 2

* Hide unused parameter

* Only enable UpSampling2D since the others are not currently supported by TVM

* Improve error messages and code layout

* add nnvm upsampling symbol

* add upsampling mxnet frontend

* add doc for upsampling op

* cleanup upsampling test

* minor fix

* use schedule_injective for upsampling

* upgrade tvm

* add more op for onnx 1.0

* fix syntax

* fix lint

* fix

* update 1.0

* fix

* update model

* Add __declspec(dllexport) to class Node and class Op if build on Windows, so that it can be referenced in some test code in MXNet

* correct typo

* reuse the macro in c_api.h

* revert changes on base.h

* one more place to revert

* use enclosing extern c block

* revert a change caused by copy

* remove a space

* dump lowered ir when debug log

* avoid calling tvm.lower() twice when not debug

f9ed337552a26cc55c7c0fb22cd54839ee13f19c added a call to std::log
without including cmath

At present, the build system assumes that the Unix archiv utility 'ar'
will always be present in PATH, but this might not be true if a separate
toolchain is being used. This patch makes the build system a bit more
cross-compile capable.

* Add creating gradient symbol

* Fix lint

* Address comments

* Fix typo

* Address comment

* Update docs

* Add backward operator to enable backward graph

* Fix testing

* Refactor top level1 test code

* Fix format

* Test

* Added zeros ones op

* Register fill_like operator

* Fix unit test

* Add compile library tutorial

* Clean output

* Refactor with sphinx gallery

* Refactor

* Change title and other minor fixes

* keras frontend tutorial

* fix

* vgg16 success

* remove six.PY2, use sys.version_info;
convert_activation() accepts activation type name(str, e.g. 'relu') as input;

* add convert_merge

* fix convert_batchnorm;
improve tests

* fix lint

* add numpy-style pad operator

* deal with asymmetry padding

* resnet50 success

* fix pool_convert; xception passes test

* update tvm

* fix bias error; all tests pass

* use > >, not >>

* [SYMBOL] Add __iter__ and GetChildren for symbol

* [SYMBOL] Fix lint

* Symbol __getitem__ using list_outputs() is too expensive, when it only cares about the output count in most cases

* Add cython cmake

* GetNumOutputs() and __len__ changes per PR comments

* set commit for tvm

* fix mxnet multi outputs

* add test case for multi outputs

* fix

* fix

* fix

* fix index

* use json hack

* fix test cases

* fix test cases

* fix test cases

* fix test cases

* fix test cases

* fix test cases

https://apple.github.io/coremltools/generated/coremltools.models.neural_network.html?highlight=permute#coremltools.models.neural_network.NeuralNetworkBuilder.add_permute

* [CMPL] Add Support for Other Data Types

* [CMPL] Add test

* [CMPL] Fix

Variable declaration correction.

* Make sure to install the nnvm_compiler library.

* Add conda packge.