* half2

* add epsilon

* [Paddle Inference]support n lookup_tables fuse to embeddinglayernorm(3)

Co-authored-by: NLi Min <11663212+limin2021@users.noreply.github.com>
Co-authored-by: Zhou Wei <1183042833@qq.com>
Co-authored-by: littletomatodonkey <dazhiningsibuqu@163.com>
Co-authored-by: Nzhupengyang <1165938320@qq.com>
Co-authored-by: NSiming Dai <908660116@qq.com>
Co-authored-by: Njiangcheng <thisjiang@qq.com>
Co-authored-by: NGuanghua Yu <742925032@qq.com>
Co-authored-by: NTeFeng Chen <ctfeng66@163.com>
Co-authored-by: NZhong Hui <zhonghui.net@gmail.com>
Co-authored-by: Nliu zhengxi <380185688@qq.com>
Co-authored-by: NJason <928090362@qq.com>
Co-authored-by: NDingjiawei <327396238@qq.com>
Co-authored-by: NChen Weihang <chenweihang@baidu.com>

* add readme for phi

* polish details, test=document_fix

* refine gil use

* [CodeStyle] remove all future import

* revert test_error.py

* restore future import in example code

* - Conv grad changed for MD

* - lint

* - compilation fix

* yet another lint

* fix libpaddle soname mismatch error

* fix windows failed

* polish linux and windows make impl

* unify winddows lib name

* fix windows error

* revert copy dst change

* revert naming change

* revert windows change

* fix gpups compile failed

* it is a test for my idea,test=coverage

* it is a test,test=coverage

* it is a test,test=coverage

* optimizing coverage ci for build develop twice

* it is a test to delete a unnitest,test=coverage

* optimizing pr ci coverage

* [Eager] math op sink to Cpp level

* fix ci errors

* draft version


* support + and - operator under cpp directly

* add static test

* polish code

* promote types or unify right type to left

* recover static test case

* polish code and fix some ci errors

* support complex and polish code

* fix conflicts

* fix windows ci errors

* fix windows-inference-ci errors

* polish and fix tests

* fix test case

* polish code

* [Eager, Performance optimization] support multiply( * operator) to sink to Cpp layer

* rm useless glog

* polish code

* polish code and fix code-format

* polish code

* fix CI

* polish code

* Support both use_calc_stream and sync_op in send recv APIs (#46023)

* add batch_norm prim2orig rule
Co-authored-by: NWen Sun <35923278+HermitSun@users.noreply.github.com>

[Eager, Performance optimization] math op sink to Cpp level ( + and - operator as an example ) (#45811)

* [Eager] math op sink to Cpp level

* fix ci errors

* draft version

* draft version

* draft version

* support + and - operator under cpp directly

* add static test

* polish code

* promote types or unify right type to left

* recover static test case

* polish code and fix some ci errors

* support complex and polish code

* fix conflicts

* fix windows ci errors

* fix windows-inference-ci errors

* polish and fix tests

* fix test case

* polish code

* polish code

* polish code and fix code-format

* polish code

* polish code

* polish code

* polish code

Signed-off-by: NZhang Jun <jzhang533@gmail.com>
Signed-off-by: NZhang Jun <jzhang533@gmail.com>

* enable eager mode on xpu, test=kunlun

* add numpy support to xpu

* fix tensor using error

* fix  error, test=kunlun

* fix failed tests, test=kunlun

* support rsqrt_p

* refine code and ut

* add_prim_rsqrt

* fix ut

* open the clip_extra flag in paddle.static.save_inference_model, test=allcase

* set the defalut value of clip_extra in TracedLayer from False to True, test=allcase

*  Disable TF32 to solve accuracy for test_trt_conv_pass and test_trt_deformable_conv in A10 GPU.

* Add ExecutionConfig and fix last-live-op bug for standalone executor

* Improve code design