* Fix bug of test_sync_batch_norm_op_static_build accuracy problem under
cuda12.

* Remove useless code modification.

* Remove useless code modification.

* Add PHI option in cmake (#54462)

* add phi option in cmake

* modify default option

* fix windows bugs

* perfect log

* change default mode

* change phi so on for py3

* fix mac bugs

* fix mac bugs

* delete code

[inference][trt]unary, bitwise_not,one_hot op and inspector ci adjust for TensorRT8.6 (#54251) (#54579)

* unary bitwise_not adapter tensorRT8.6 in Paddle-TensorRT

* one_hot_op ci fix

* Update test_trt_inspector.py

delete log

[cherry-pick 2.5][Zero-Dim] paddle.nanmedian/count_nonzero/logspace support 0D, add some 0D case (#54649)

* [Zero-Dim] add 0D test case (#54581)

* [Zero-Dim] paddle.nanmedian/nanquantile support 0D Tensor (#54500)

* [Zero-Dim] paddle.nanmedian support 0D Tensor

* fix CI

AMP 动静态图黑名单统一

* [AMP Prim OP]support some prim ops for bf16 dtype part3 (#54368)

* support some prim ops bf16 dtype

* fix cmake

* [AMP Prim OP]support some prim ops for bf16 dtype part5 (#54422)

* support some prim ops for bf16 dtype

* remove useless code

* support some prim ops bf16 dtype (#54399)

Fix fused gemm bugs

[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>

* [Distributed] Add pipeline opt memory (#54505)

* fix bug of release output in pp

* cherry-pick 2.5 Numpy version is too high, resulting in issues with single testing

* fix

* replace requests with httpx

* set timeout=3

* replace requests with httpx

* replace request with httpx

* test

* repalce requests with httpx

* test

* replace requests with httpx

* replace requests with httpx

* modify paddle_build.sh

* fix bug

* [CINN] reopen mean/cumsum/instance_norm op's prim+CINN test

* remove repeat test_mean_op in cmake

* refactor scalar attribute

* fix op build

* fix merge conflicts

* fix coverage ci

* add ldexp api

* fix ldexp

* Update math.py

fix math.py

* rewrite the ldexp

* Simplify ldexp implementation

* fix codestyle

* Update math.py

* Update math.py

* modify the test_ldexp.py

* fix the input bug of np.ldexp

* fix the bug of np.ldexp in windows

* modify the ldexp function and add the dtype check of output

* Update test_ldexp.py

* fix the dtype

* fix codestyle

* Update python/paddle/tensor/math.py

add note for code description
Co-authored-by: Nzachary sun <70642955+sunzhongkai588@users.noreply.github.com>

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

* fix_bug

* fix bug

* update roll convert

* [AutoParallel] fix sharding for 0D tensor and amp-o1

* add amp for sharding unittest

* Fix compilation error by using thrust

* fix

* update

* update

* update

* update

* update tag

* fix cutlass

* update cub tag

* update

* update