* add FP16 dtype for CastNumpy2Scalar

* fix throw message

* add test

* fix SyntaxWarning

* test skip for float16

* fix dtype mistakes

* fix div 0 error of NoamDecay

* add unittest

* Update lr.py

* [CustomDevice] add custom device api

* update

* update

* test=document_fix

* update

* update

* add  examples

* [KUNLUN] rename test_pool_max_op.py

* update xpu toolchain

* fix incorrect output shape of broadcast

* add unittest

* fix mod 0 error

* fix div 0 error in floormod

* remove fluid.ir.RegisterPassHelper PassDesc and RegisterPass

* proto import problems

* change import way of pass_desc_pb2

* change sys.path

* change the way of import framwork_pb2

* add fluid_path directory from path.dirname

* fluid_path changed

* support 0d tensor for interpolate

* support 0d tensor for interpolate

* add xpu unittest for interp

* update unittest for interpolate

* fix coverage

* fix code style

* fix for coverage

* fix coverage

* support inplaced variable in cinn_launch

* fix error hint when compiling

* fix inplaced output variable of the subgraph

* skip CinnCompiler check

* using existed definition

* fix namespace reference error

* modify error message

* update cinn tage

* fix namespace

* skip enforce check

* fix unittest attribute throw

* decouple phi custom_op

* decouple phi custom_op, remove codes

* delete custom symbol of inference

* add channel_num check for paddle.static.nn.batch_norm

* fix bugs

* fix bugs

* add dimension check for interpolate

* modify dimension check for interpolate

* add unittest to size check for interpolate

* fix incorrect shape check for interpolate

* split size check and add unittests

* add axis check in UniqueRawInferMeta

* add unittest for negative axis

* simplify check for unique

* add elements count check in atan2

* add unittest and pre-check in inferMeta

* add dimension check

* fix the div 0 errors in psroi_pool

* fix case 7

* rool back sth.

* add zero size check in matrix_power_kernel_impl.h

* add zero size check in matrix_power_kernel_impl.h

* add zero size check in unittest

* bug_fix

* bug_fix

* bug_fix

* bug_fix

* bug_fix

* bug fix

* bug_fix

* bug_fix

* add static check

* delete the dy codes

* support negative index in repeat_interleave

* add unittest

* Unify the gpu implementation of stack and unstack to reuse the optimization.

* Optimize the cuda implementation of unstack.

* Use GpuMemcpyAsync instead of memory::Copy.

* Fix error of calculating the index.

* Use FastDivMod to further imporve the performance of unstack.

* change int -> int64

* Update python/paddle/fluid/tests/unittests/ir/inference/test_trt_convert_one_hot.py

* update erf gumbel_softmax ..ops

* lint

* reset sequence_conv

* reset exponetial&interp nearest

* rm flags_retain grad in pybind

* retain grads for xpu test

* set retain grad for xpu

* rm flag

* lint

---------
Co-authored-by: Nwanghuancoder <wanghuan29@baidu.com>