* [MLU]add sync stream ops and broadcast pytest

* [MLU]fix broadcast pytest to add data type

* merge legacy to fluid

* Remove legacy code

* Remove legacy code

* Remove DataType test

* Using Tensor directly instead of using EagerTensor

* support gradient_accumulation

* make test_imperative_lod_tensor_to_selected_rows longer

* make test_imperative_lod_tensor_to_selected_rows longer

* refine code

* Rename all EagerTensor to Tensor

* Rename some EagerTensor to Tensor

* rename EagerTensor to EagerVariable

* add more test

* fix different device gradient_accmulator bug

* merge develop

* remove useless tests

* [Eager] Support GradientHook before running seperate GradNode

* Fix CI issue

* Fix CI issue

* pten matmul cuda kernel support bf16

* avoid custom kernel deps on pten_function_api

* Revert "pten matmul cuda kernel support bf16"

This reverts commit 5d520845b9a189375677276efb673235ed8e5ee0.

* refine code

* fix compile

* fix test_split_api

* move bernoulli kernel to pten

* follow comments

* relocate code of interpretercore gc

* fix and add more tests

* remove unwanted changes

* check only concat and elementwise

* move check to a function

* add todo comment

* Revert "fix ptq fc attr name fuse_activation->activation_type"

This reverts commit ffd023353a5e9b0fd15e81b9e9f9fe1794035017.

* save the name lists of the input,internal and output variables of a subgraph as its attribute

* fix compile error

* move trunc to pten

* modify according to YuanRisheng's comment

* clean useless header in pten core

* fix compiled failed

* fix cmake target

* fix typo

* resolve conflict

* add softplus op for kunlun2. test=kunlun

* add softplus op for kunlun2. test=kunlun

* fix code style. test=kunlun

* fix code style. test=kunlun

* add more test cases. test=kunlun

* first commit

* teller fix

* bug fix

* enable for pool2d only

* fix global_pooling issue

* pooling_type

* fix test

* remove register of matmul_v2 kernel

* delete matmul_v2 grad register in fluid

* update kunlun label_smooth unitest
*test=kunlun

* minor
*test=kunlun

* update generate_pd_op_dialect_from_paddle_op_maker.py

* update mlir tensor load interface

* refine

* fix bug

* fix

* refine

* fix

* 3

* fix

* codestyle
Co-authored-by: weishengying <1343838695@qq.com>

* remove fluid device context deps

* fix compile failde

* update inference ut to support nhwc format

* update ut and pass OpCompat

* update ut

* update ut

* Update comment

* [Eager] Support test_imperative_hook_for_layer with _test_eager_guard()

* Polish code name style

* Fix a error name

* Polish code, make it clear and simple

* change ci using mutable_data() check's directions from paddle/pten to paddle/pten/kernels

* change echo info from paddle/pten to paddle/pten/kernels

* move lerp to pten

* refine include

* move files

* refine code

* pten matmul cuda kernel support bf16

* fix pten kernel name

* add matmul_grad bf16 kernel

* add emptylike bf16 kernel

* fix compile

* suppport rocm

* fix error

* fix rocm

* add bf16 header file

* fix compile

[Paddle-Inference] support preln-ernie: add preln_emb_eltwise_layernorm_op, preln_skip_layernorm_op (#39570)

* support preln_ernie: add preln_emb_eltwise_layernorm_op, preln_skip_layernorm_op

* support preln_ernie: add preln_emb_eltwise_layernorm_op, preln_skip_layernorm_op

* merge legacy to fluid

* Remove legacy code

* Remove legacy code

* Remove DataType test

* Using Tensor directly instead of using EagerTensor

* support gradient_accumulation

* make test_imperative_lod_tensor_to_selected_rows longer

* make test_imperative_lod_tensor_to_selected_rows longer

* refine code

* Rename all EagerTensor to Tensor

* Rename some EagerTensor to Tensor

* rename EagerTensor to EagerVariable

* add more test

* merge develop and refine code