Cherry-pick of #41521

[cherrypick-2.3] modify infer gpu memory strategy (#41427), remove cudnn_deterministic=True (#41341)  (#41491)
Co-authored-by: NJingZhuangzhuang <75348594+JZZ-NOTE@users.noreply.github.com>

* add one_hot gpu hint

* move allow_out_of_range judgement

* delete useless unittest

Cherry-pick-PR41407, fix device_id bug for final_state op in multiprocess testcase (#41407) (#41475)

Cherry-pick PR #41407

[Cherry-Pick]Fix eager try catch (#41438)

【chery-pick #41430】fix bug of random compile failure, due to incorrect compile order of dependencies

cherry-pick

fix compile bug of windows cuda11.5 #41433

* add ipu uts

* fix ut

* split PR

* fix ut

* rm ut

* remove paddle_ipu shared library

* fix unique_name

* [Eager] Support test_layers's test cases switch to eager mode

* Update batch_norm _C_ops action to fix CI

* Use None instead of new EmptyTensor

* Updated var name

* Make sure to switch eager mode, Fix Coverage_CI

* Remove _non_static_mode statement

* Remove batch_norm dispensable input statement

* Polish batch_norm code

* Fix CI issue

* Update __init__.py

* Update math.py

* Create test_nanmean_api.py

* Update __init__.py

* Update __init__.py

* Update math.py

* Update test_nanmean_api.py

* Update __init__.py

* Update math.py

* Update test_nanmean_api.py

* Update test_nanmean_api.py

* Update test_nanmean_api.py

* Update math.py

* Update test_nanmean_api.py

* Update math.py

Update the nanmean example code

* Update math.py

* Update math.py

* Update math.py

Remove redundant code in nanmean

* Update math.py

change default keepdim = False

* Update test_nanmean_api.py

add nan into self.x

* Update test_nanmean_api.py

rerun CI check

* Update test_nanmean_api.py

* update code of nanmean in python/paddle/tensor/math.py and test_nanmean_api.py

* Update test_nanmean_api.py

update code format

* Update test_nanmean_api.py

update code format

* Update test_nanmean_api.py

add check grad code.

* Update math.py

update nanmean's describe of Args x

* Update test_nanmean_api.py

update format and release the test_case(self.x, keepdim=True) in check grad code.

* Update test_nanmean_api.py

Update gradient checking method

* Update test_nanmean_api.py

update code format

* Update test_nanmean_api.py 

Update code format and copyright in test_nanmean_api.py

* Update math.py

update arguments describe and code example

* Update math.py

修改了nanmean的axis参数的文档描述。

* Update math.py

updata nanmean's sample code (:name: code-example1)

* Update math.py

修改nanmean的example code 错误

* Update math.py

update example code

* Update math.py

update example code of nanmean

* add real and imag yaml

* add roi_align and roi_pool yaml

* add qr yaml

* add psroi_pool yaml

* fix bug

* fix param bug of psroi_pool

* fix infrt problem

* fix merge bug

* update

* add conv yaml

* add backward

* remove useless code

* fix bug

* fix bug

* revert fluid dygraph conv2d

* remove useless infermeta function

* fix meta fn deluplicat error

* conv using custom impl

* remove amp include

* fix bug

* use cudnn = true

* fix test mkldnn caching bug

* fix less than

* fix some tests

* fix additional 3 unittest case

* fix bug of data transform in inference executor

* fix bug

* move yaml

* add momentum yaml

* delete code

* delete some code

* add meshgrid backward

* delete code

* fix compile bugs

* update name

* update name

* fix test

* fix fleet bind

* update name

* update name

* fix test

* fix gpups wrapper

* remove Push/Pull/Load/Save with context in client and wrapper base class

* fix

* fix

* remove some interface

* fix

* remove

* code style

* recover

* fix

* remove code unused

* remove some unused table & accessor & CommonDenseTable => MemoryDenseTable

* fix

* fix

* fix

* recover

* remove unused code
Co-authored-by: Nesythan <esythan@126.com>

* add fake index and unittest for multiclass_nms3 trt

* modify unittest

* [Refactor] refactored eager_gen.py PR #2

* [DoubleGrad PR #1] Decoupled code generation logics for Dygraph ForwardFunctions and GradNodes

* Fixed minor issue

* Adjusted logics of GenerateNodeCreationCodes and GenerateForwardDefinition

* Fixed issues

* Supported higher-order grad node generation

* [DoubleGrad PR #4] Supported higher-order GradNode generation

* [DoubleGrad #4] Bug Fixes to Double Grad Node Generation

* Fixed yaml typo

* Fixed yaml typo

* fixed minor issues

* [DoubleGrad PR #5] Enabled gradient computations for grad_tensors passed to paddle.grad()

* Fixed minor issue

* Fixed CI-Inference issue

* Fixed CI-inference issues

* [DoubleGrad PR #7] paddle.grad() to copy backward graph before backward run

* Fixed minor issues

* Fixed issue with backward graph construction logic

* Fixed implementation issues with backward graph reconstruction

* Fixed unittest issue

* Fixed issues

* [DoubleGrad PR #8] Enabled triple grads for sigmoid and matmul

* Fixed issues with phi kernel

* Added triple grad test case

* Fixed minor issue

* add roi_align yaml

* fix bug

* support process group in dp with fleet api

* update

* fix uts

* update

* add nms op and batched_nms api

* switch autotune

* implement AutoTuneCache

* implement AutoTuneCache class

* add pybind api

* add dygraph test

* support static mode and eager mode and improve unittests

* rename the SwitchAutoTune Class and improve tests

* improve AutoTuneStatus and reduce the cost of tests

* eager math op, test=develop

* eager support lookahead, test=develop

* refine,test=develop

* refine doc, test=develop

* refine,test =develop

* refie, test=develop

* refie, test=develop

* refie, test=develop

* test_paddle_multiprocessing

* refine, test=develop

* refine, test=develop

* fix bug, test=develop

* refine, test=develop

* dataloader, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* test_datasets timeout, test=develop

* refine, test=develop

* [Eager] Fix empty tensor Initializer bug with shape=[]

* [Eager] Fix empty tensor Initializer bug with shape=[]

* ignore two unittest

* fix unittest

* enable new executor by default

* enable stream safe allocator

* test=document_fix;test=coverage

* do not use scope in op kernel

* fit empty program for new executor

* fix communication depend

* fix test_sync_batch_norm

* skip unsupported place

* refine datatransfer

* fit for dirtributed program

* fix dependencpy

* fix some ut

* add stack yaml

* add stack yaml

* add stack yaml

* add no_need_buffer

* refine no_need_buffer declare

* remove original grad infershape

* revert stack op

* support inplace no_need_buffer

* fix

* use padle.add