1. Allocate memory for output before compute.
2. Rename 'ctc_decode' to 'ctc_align'

1. Remove 'top 1'(or argmax) from CPU and GPU kernel
2. Add a new test case
3. Refine doc

* fix while_grad_op first step loss lod problem

* optimize code

* "unified operators"

* "add CUDNN register"

* "add use cudnn attribute"

* "add attribute"

* "test conv tranpose op"

* "remove duplicated attr"

* "fix op test"

* "add attribute to set cudnn"

* "add more log"

* "need layout op register support"

* "add more log"

* "change GetExpectedKernelType "

* "fix Get attr in conv_op"

* "fix CI"

* "fix tests"

* "removed kernel priority fallback"

* "fix CI"

* "fix stack pointer bug"

* "refine buggy interface"

* "add const cast to save life"

* "fix get_output_with_grad"

* "fix op test with dataformat"

* ""fix pooling

* "fix pooling test"

* "fix CI"

* "fix with_gpu error"

* "add transform needed functional check"

* "fix unpack list error"

* "comment out parallel.do temporary"

* "fix CI"

* "fix compile doc error"

* "make threshold larger"

* update code comments.

* update the comments.

* follow comments.

We need this operator to assign value to a tensor and the values are stored in the program so that they can be used independent of python.

Async GRPC sendrecv

feature/parallel_gpu

* add lod tensor ToAbsOffset test

* add share lod to topk op and softmax op

* test dist word2vec

* multiple trainers work

* Add Seq2BatchFunctor, which will be used in WarpCTCOp.

* Implement WrapCTCFunctor and WrapCTCKernel.

* Add unittest of warpctc_op.

* Modify the check_output inferface in python unittest framework to allow check a subset of outputs.

* Use absolute offset lod in warpctc_op and related functors.

* Refine the comments of warpctc_op.

* The new python unittest supports checking a subset of the outputs, so revoke the previous change.

* Rename the transform from LoDTensor to Tensor with shape [max_sequence_length, num_sequences, sequence_width] to PaddingSequenceFunctor.

* Update to the newest codes.

* Rename the PaddingSequenceFunctor to PaddingLoDTensorFunctor and remove the computation of dimensions out of the functos.

* Rename Tensor::CopyFrom to Tensor::Copy

* Fix CI

* Fix compile

* create tensor in recv op

* static global function to global function

* add rename guard

* add device_data_transform

* add device_data_transform_test

* modify GetExpectedKernelType

* update operator.run

* support test test_label_semantic_roles

* optimize code

* optimize code

* rename GetActualKernelType to GetExpectedKernelType

* fix chunk_eval_op and device_data_transform_test

* add is_same_place to place

* optimize code, refine rename_guard

* refine rename guard, add GetKernelTypeForVar

* optimize code

* add some log

* rename guard

* use sub scope to create var

* fix compile

* add IsInitialized for Tensor

* add VarIsTensor

* fix op_registry_test

* test

* tmp disable priority

* restore switch_kernel.md

* code clean

This can make it easier to locate error.

* "add c++ side kernel selection"

* "add multiple kernel op test"

* "kernel selection only support cudnn"

* "better formatter"

* "small fix with UseCPU"

* "depends on change interface Get(Place, Library)"

* "fix CI"

* "fix python cudnn test"

* "leave the register cudnn op to another PR"

* "fix CI"

* "use all kernel by default"

* "fix CI"