* add matmul doublegrad op

* fix compile errors

* modify code according to review

* delete float16

* fix pool3d bug, test=develop

* fix unitest, test=develop

* fix test and fix pool2d bug, test=develop

API(lstm_unit, lstmp, sequence_mask, sequence_enumerate, sequence_conv) error message enhancement (#27572)

* API(Compute) error message enhancement on line 44, 50, 53.

* lstm_unit error message enhancement.
lstmp error message enhancement.
sequence_conv error message enhencement.
sequence_enumerate error message enhencement.
sequence_mask error message enhencement.

* Update lstm_unit_op.cc

* Update lstm_unit_op.h

* error msg enhancement.

* Update sequence_conv_op.cc

* Update lstm_unit_op.cc

* Update sequence_conv_op.cc

* Update sequence_enumerate_op.cc

* Update sequence_enumerate_op.cu

* Update sequence_enumerate_op.h

* Update sequence_pool_op.h

* error message enhencement.

* error message enhancement.

* Optimize the error message for OPs.

* Optimize the error message for OPs in details.

* fix bilateral bug

Polish the error message of fc, fused_fc_elementwise_layernorm and fused_embedding_seq_pool. (#27692)

* Polish the error message of fc_op.

* Polish the error message of fused_fc_elementwise_layer_norm op.

* Polish an error message in fused_embedding_seq_pool_op.

* - condidate fix to issue #25537

test=develop

* - UT for transpose NHWC

test=develop

test=develop

* fix distributed error info

* bug fix; notest

* error info refine

* update error info

* update error info

* update error info

* bug fix

* bug fix

* bug fix

* bug fix

* add multinomial cpu kernel

* fix C++ notype error

* fix windows ci array len error

* let array len be const

* change array to vector

* add cuda kernrl with num_distribution is 1, and not support replacement=False

* add multinomial python api

* support num_distribution different multinomial distributions

* add multinomial python api unittest

* change output dtype to int64

* fix coverage prob

* optimize format

* fix dtype of output error, should be int64_t

* test=develop, optimize geo communicator 

* unique op for cuda is added

* add support for cuda

* Add cuda support for unique op.

* Add support for int32_t and int64_t.

* For old version, process by cpu

* Add VisitDataType for thrust

 optimize the error meesage for detetion_map_op

* Fix clip_by_norm_op error message. test=develop

* test=develop

* test=develop

* enhance array_to_lod_tensor_op lod_tensor_to_array_op errors information. test=develop

* iCafe 9009: Optimize the error message of OP.

* Optimize the error message of GatherTreeOP.

* fix the speed&memory of matmul

* fix the comment

* fix the memory copy

* fix the windows ci

* fix error

* fix error

* fix error

* merge develop

* enhance error messages of lookup_tale, merge_ids, data_norm

* fix

* fix error msg in .cu

* Error message opt, test=develop

* solve comments, test=develop

* fix typo, test=develop

* add float64 input to ctc_loss

* modified error message of  warpctc

* update repo and tag of warpctc

* add test for warpctc with float64 input

* modified warpctc.cmake to make sure build always

* resolved sample code bug of warpctc

* add core.ops in warpctc dygraph

* fix a bug of test

* support elementwise add, activation, matmul on Baidu Kunlun
* test=kunlun

* minor
* test=kunlun

* reconstuct the xpu directory
* test=kunlun

* minor
* test=kunlun

* minor
* test=kunlun

* minor
* test=kunlun

* minor
* test=kunlun

* minor
* test=kunlun

register log double grad kernel for cpu and cuda

* fix sgd/momentum/dpsgd/rmsprop error message

* fix test_dist_fleet_heter_ctr & peformance update