[inference][cherrypick] Implement layer_norm op using INormalization Layer and conv_fusion support bias's rank equal to input's rank  (#54590)

* [inference]conv_fusion support bias's rank equal to input's rank (#54477)

* support bias's rank equal to input's rank

* [inference][trt]layer_norm op with dynamic shape support INormalizationLayer in TRT8.6 (#54379)

* layer_norm op with dynamic shape support INormalizationLayer in TRT8.6

* Using trt layer to make layers_norm op in lower than trt8.6
layer_norm op with dynamic shape support INormalizationLayer in TRT8.6

---------
Co-authored-by: Nbukejiyu <52310069+bukejiyu@users.noreply.github.com>

* Fix compilation error by using thrust

* fix

* feat: added polygamma init code

* feat: added polygamma unittest code

* test: added more test cases

* refactor: added forward impl

* refactor: added backward impl

* test: updated cases

* refactor: updated test cases

* refactor: added more case and fixed some bugs

* test: updated ref func

* refactor: updated code style

* refactor: move the code

* refactor: updated test

* refactor: updated test

* docs: updated en doc
Co-authored-by: Nzachary sun <70642955+sunzhongkai588@users.noreply.github.com>

* docs: updated math eq

---------
Co-authored-by: Nzachary sun <70642955+sunzhongkai588@users.noreply.github.com>

Support code generation for op conv2d_transpose, conv3d_transpose, depthwise_conv2d_transpose (#54242)

* optimize logsumexp in small data scale

* fix

* fix

* add #pragma once

* swith to use aligned_vector and support arbitrarily shape

* fix store

* fix store

* refine for special cases

* try

* fix

* update

* fix

* fix all_reduce

* try

* fix rocm bug

* fix rocm bug

* fix rocm bug

* fix rocm bug

* fix rocm bug

* fix rocm bug

* fix rocm bug

* fix rocm bug

* add fp&bf16 bernoulli

* add check_dtype & fix error

* fix rocm error

* floor div support float/double/bfloat16/float16

* add ut

* fix bug

* fix fft.ifftshift for floor_divide upgrade

* fix comment

* fix bugs

* fix bug

* Optimize perf of broadcast matmul

* support more dtype

* support master_grad for adam and momentum

Co-authored-by: zhangting_2017@163.com <zhangting2020>

* static graph autogen code for shape op

* fix onednn

* fix onednn

* [ROCM] fix multihead_matmul

* skip bf16 uts

* update

* update

* update cmake

* update

* update

* update

* update

* Revert "update cmake"

This reverts commit 1e1dc1b2bc9967b725201272607f939260070fd4.

* update

* update

* update

* update

* support activation prim op bf16 dtype

* remove useless code

* update_c++17

* update_c++17

* fix windows bug

* solve cirle depend

* solve cirle depend

* solve cirle depend

* solve cirle depend

* solve cirle depend

* fix windows bug

* fix compiler error

* fix compiler error

* update eigen3

* update eigen3

* update eigen3

* fix mac-py3 compiler error

* update C++17

* fix mac compiler error

* fix compile error

* fix coverage_compiler error

* fix coverage_ci_problem

* fix coverage_error

* fix_kunlun200 compile error

* fix kunlun200 compiler error

* fix compile error

* fix compiler error

* fix py3 failed test

* fix kunlun200 compiler error

* test

* fix test error

* fix test error

* fix test error

* test

* test

* fix mac py3 error

* fix mac py3 error

* fix mac py3 error

* fix test error

* fix test error

* fix compile error

* fix compile error

* fix compile error

* test

* test

* fix compiler error

* test

* test

* debug on ci

* fix compiler error

* fix compiler error

* test

* fix cinn compiler error

* test

* fix rocm cmpile error

* fix cinn and kunlun compile error

* update c++14

* Update flags.cmake

* softmax fwd: force vec size to 1 when dtype is float

* use 1024 as threshold to use cudnn

* Reimplement the check_nan_inf function as check_numerics kernel.

* Remove the cpu implemention to phi.

* Add ifdef for the including of omp.h.

* Move the use of FLAGS_check_nan_inf_level out of header file.

* Implement a common PrintAndThrowError function.

* Fix the error using of __NVCC__, which should be instead with __CUDA_ARCH__.

* Add dependency of phi.

* Polish codes and unittest.

* create phi so

* fix ci bugs

* fix py3 bugs

* add file

* fix py3 bugs

* fix windows bugs

* perfect so

* fix py3 bugs

* delete all static target in phi

* fix windows bugs

* fix py3 bugs

* fix ci bugs

* fix windows bugs

* fix bugs: gflags can't be linked by dynamic and static lib

* fix bugs that can not load 3rd party

* fix ci bugs

* fix compile bugs

* fix py3 bugs

* fix conflict

* fix xpu bugs

* fix mac compile bugs

* fix psgpu bugs

* fix inference failed

* deal with conflict

* fix LIBRARY_PATH bug

* fix windows bugs

* fix onednn error

* fix windows compile bugs

* fix windows compile bugs

* fix test_cuda_graph_static_mode_error aborted

* fix windows bugs

* fix mac-python3 error

* fix hip compile bugs

* change mode to static

* change to static mode

* fix ci bugs

* fix py3 bugs

* fix windows bugs

* fix bugs

* add static flag

* add PADDLE_API

* change position of PADDLE_API

* fix windows bugs

* change mode to dynamic lib

* fix windows static bugs

* deal with conflict

* fix windows unit bug

* fix coverage

* deal with conflict

* fix windows-inference

* fix py3 bugs

* fix bugs when compile type_info

* fix compile bugs

* fix py3 bugs

* fix windows bugs

* fix windows openblas

* fix xpu bugs

* fix enforce_test in windows

* update code according comment

* fix windows cmake bug

* fix windows bugs

* fix windows bugs

* delete cinn unittest

* fix cinn bugs

---------
Co-authored-by: lzydev <1528794076@qq.com>

* add log for memory stats

* fix string_split in einsum

* Try to increase the repeat of autotune and fix the setting of allow_tf32_cublas.

* Change the repeat of cublaslt to 10.

* Use FLAGS_cublaslt_exhaustive_search_times as repeats.

* Fix compiling error on CI.

* Polish the key and simplify codes.

* move raw kernels to legacy

* Update elementwise_add_kernel.cu

* fix