* Integrate APU

* fix code style issue

1. Auto transformation between data format, data type and memory type.
2. Add device placement optimization strategy.

add support splite with num_splits=1

1. Support int32 input data type.
2. Support GatherV2 op
3. Add transpose to ExpandDim op.

1. add capability show information for mace_run
2. remove flatten_atrous_conv of tensorflow transform_graph

1. input_data_formats flag in yaml supports OIHW
2. MaceTensor support OIHW data format

* Added floordiv() operator

* Restored original indentation

* Update eltwise.cl

Update eltwise.cl according to the recent fix.

* Support LeakyRelu conversion from tensorflow

* Code style PEP8

add onnx_optimizer tool and update onnx info in docs

1. Abstact android and arm linux to one format
2. Support cross compilation for ARM linux
3. Related issue #36

1. Merge the Operation and OpKernel into Operation.
2. Add OpDefRegistry for registering the definitions of Operations.
3. Add information for MaceStatus.
4. Unify 'BatchNorm' and 'FoldedBatchNorm' into 'BatchNorm'
5. Remove unused op 'Proposal'

* add friendly check for batch normalization layer

* coding style

* reduce the too long line length

This reverts merge request !748