* support fp16 for maxout op

* format code

* change api

* add test for static float16

* format code

* formatting code

* atol alignment

* experiment—1

* experiment-2

* experiment-3

* format code

* trans fused_feedward Compute function to phi

* add register info

* remove maxfunctor

* move fused feedward to phi

* remove sig file

* remove fliud include

* add include

* add include

* add sig file

* add output register info

* fix sig file

* Update fused_feedforward_sig.cc

* fix grad kernel

* update output register info

* fix

* open fused_feedforward static build

* add optional and fix code style

* fix output info for fused attention

* add optional param

* merge

* modify concat_grad add sum comp rule

* modify opcompat

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* [XPU] remove scale_loss in parallel.py

* [XPU] throw Unimplemented when using Reducer

* add pad op

* add_some_code

* modify some code

* add some code

* add some code

* modify some code

* add some code

* modify some code

* Update composite_backward_api.h

* modify some code

* add some code

* add some code

* add some code

* add scatter_nd_add comp

* add scatter_nd_add prim

* fix

* fix

* add public_python_api in TestScatterNdAddSimpleOp setup function

* fix composite_backward_api.h

* fix composite_backward

* add test cases

* fix composite_backward_api.h, unittest

* Fix fused_attention_op and fused_feedforward_op bugs in xpu

* Fix d_x alloc errors for fused_feedforward_grad_kernel

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* test,test=develop

* optimize embedding deterministic mode

* fix compile error

* change FLAGS_cudnn_deterministic to int64

* fix 700 error

* add ut

* fix ut

* fix ut

* fix win32 ci

* fix flags with PHI_DEFINE_EXPORTED_int64

[Bug Fixs] fix bugs when using cast<int64_t, int32_t> in xpu/cross_entropy kernels, *test=kunlun (#53325)

* Optimize prompt information

* add_information

* add_information

* support register single .cu file

* add register GPU kernel function

* create KernelMinMax to optimize the performance of histogram op in GPU

* change to block and warp wise operation

* remove the time in DtoH

* fix a bug

* add flags for phi

* fix compile bugs

* fix ci bugs

* fix inference bugs

* fix cinn' bugs

* fix cinn bugs

* perfect code according comment

* fix ci bugs

* fix ci bugs

* fix shared memory over usage in embedding grad kernel on determistic mode

* use IdT as interger dtype

* add fp_bf for pool_max_withidx

* fix some error

* fix error

* codestyle error

* fix masktype

* fix input bf type

* input bf dtype convert error

* back to convert input to bf16 first

* fix convert error

* fix bf16 grad check

* support 0d tensor for shape and squeeze onednn kernel

* set python api for shape op ut

* Fix the calculation of layer_norm_bwd

* fix

* add 0D support for trace, test=allcase

* fix trace gpu kernel 0d error, test=allcase

* fix windows error, test=allcase

Add paddle.geometric.weighted_sample_neighbors API