* Implement slice grad operator. #8130
* test slice grad operator and bug fix
* Fix pre commit style

* Add slice op.

* Remove using from header file and fix doc.

* Fix doc

* Small fix.

* add basic interface for learning rate decay
* add exponential_decay
* add natural_exp_decay
* add inverse_time_decay

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`

* Support many data types of several operators

* SeqConv only support float/double

* Revert adagrad

* Adding Adadelta optimization operator
* Making inputs and outputs conform to naming convention
* Removing type alias from header files
* Fixing Adadelta documentation in comments
* Addressing code review feedback

Make implement an operator less noisy.

* a simplest SGD op