* fix softmax_with_cross_entropy cuda kernel overflow bug, test=develop

* replace old macro & for condition, test=develop

* polish details, test=develop

* add new macro BOOST_GET_SAFELY & unittests, test=develop

* add different macro type, test=develop

* fix get macro type in executor, test=develop

* four macro part change backup

* using one macro for all case, test=develop

* revert attribute change, test=develop

* change to three func to solve gcc4.8 bug, test=develop

* polish some details, test=develop

* test=develop modify some op for dyg rcnn

TemporaryAllocator is a singleton used for allocating memory for Cudnn. Since it is a singleton, we can delete it for better performance in memory.

We replace TemporaryAllocator by CUDADeviceContextAllocator and CUDADeviceContextAllocation, which uses stream callback to delete the memory allocated for the stream to avoid singleton.

Also added data_feed_proto to operator to fix CI in CPU compilation

* Add generate_mask_labels_op to support Mask-RCNN.
* Refine sigmoid_cross_entropy to support nomalize mode.
* Fix generator_proposals_label.
* Use DeviceTemporaryAllocator in roi_pool and roi_algin.
* Remove shape check in data_feeder.

Add flatten op interface and enhance APIs about detection to support variable-length image. (#12422)

* add flatten api&enhance detection api

* unify shape_op data type

* update API.spec

* Add shape op to get the shape of variable.

* Rename get_shape to shape.

* Add checker for output and fix comments.

* add data layout

* fix ci

There are mainly following fixes:

- take `DeviceContext` as the template parameter of math functors and OpKernel instead of `Place`
- remove `eigen_device` interface in base class  `DeviceContext`
- remove `GetEigenDevice` interface in `ExecutionContext` and base class `DeviceContext`
- remove unused `platform::EigenDeviceConverter`
- rename `REGISTER_OP_GPU_KERNEL` to `REGISTER_OP_CUDA_KERNEL`
- rename `USE_GPU_ONLY_OP` to `USE_CUDA_ONLY_OP`

* init batch norm op

* prepare input output

* compute mean_out var_out save_mean save_var on CPU

* active is test

* use eigen to do computation

* complete batch norm forward

* set default momentum to 0.9

* add batch norm grad op in CPU

* add tensor_format and NHWC support, add python test

* add test training

* add batch norm gradient test

* improve comment, fix foward Python UnitTest

* add gradient test

* fix eigen warning

* follow name style

* fix a bug

* change float to T

* add simple forward test

* test with different place

* add backward test

* refine python test

* remove old python test code

* code clean

* follow code style

* update comment

Make implement an operator less noisy.

* add inputs

* add ut for multiple inputs

* fix AddToLayer

* op_desc -> op_proto

* CreateArgumentOffsetMap -> CreateInOutOffsetMap

* move CreateInOutOffsetMap from OperatorBase to op registry

* arg_idxs_ -> in_out_idxs_

* Convert `op` --> `operators`
* Remove AddType in OpProtoMaker, because type is part of registry.
* Rename CPU_OR_GPU --> DEVICE_TYPE in registry macro.

* Refine register methods, make Op can get rid of whole-archieve
* `USE_OP` before a op is used.
* Add unittest for add_op.