* Add Preface

* Add demo code

* Save file

* Refine code

* seems can work

* use elementwise strategy

* Use ElementwiseComputeEx

* Add comments

* extract functions from operator

* Refine code

* Follow comment

* code refine

* add op_fuse  pass

* add backward

* code refine

* use TopologySortOperations

* follow comments

* refine IsFusible

* code enhance

* fix op_fusion_pass

* refine code

* refine fuse_elemwise_act_op

* adjust the input and output

* refine logic

* add intermediate_edge

* disable inplace

* follow comments

* refine logic

* follow comments

* Remove the removable IntermediateOut

* change strategy

* code refine

* enable fuse backward

* code refine

* code refine

* rename unit test

* follow comments

* framework cmake

* Update CMakeLists.txt

fix a few other things

* data_type

* "remove tabs"

* cudnn widndows

* "add comment"

* "windows support"

* "fix cmake error"

* add rw lock test

* optimize read_write and wirte_read test

Reduce one level of inheritence.

* "cherry picked cpp tests"

* "cherry picked"

* "cherry picked tests"

* "merge develop branch"

add doc

We can get endpoints of a reader chain.

* move to platform

* "move init from framework to platform"

* "remove used init"

* "fix ci"

* "fix ci"

* "fix generic"

* "fix ci"

* "fix ci"

* "fix ci"

* "disable fragile test"

* wip

* use executor.complete to end trainer

* fix build

* fix build with distribute off

* fix typo

* fix cmake typo

* fix build

* Use Builder Chain to decorate new builders. It is easy to extend
  builders.
* Make graphviz path as a build strategy, not a FLAGS.

* remove net op and cond_op

* fix cpplint

* fix dependency

* delete backward_test; fix compile

* disable batch_norm backward

* rm test_net.py

* make batchnorm test independent of backward.cc

* make test_layer_norm_op independent of backward.cc

* make test_layer_norm_op independent of backward.cc

* delete unused code

* clean up