* Use correct master weights in AdamW.

* Just modify the master weight.

* Update for CI Coverage.

* Additionally change to not throw error from inside pass.

* Update Xbyak and add bf16 fast performance verification

* Fix formating

* Change LOG message

* Trigger an update of a new tag

* fix split trt plugin initialize

* update

* use iwyu clean include second time, test=develop

test=document_fix ,Fix QQ group number in README

Fix QQ group number in README.

fix the broadcast for the large second input 

* try again if kunlun memory malloc failed

* minor

* support xpu inference with analysis predictor, test=develop

* merge the cmake of the xpu toolchain, test=develop

* add c-apis, test=develop

* fix a bug in extern_xpu, test=develop

As the title

* support setup.py to compile custom op

* move file into paddle.utils.cpp_extension

* support python setup.py install

* refine code style

* Enrich code and add unittest

* Polish code and api doc

* fix cpp_extension not include in package

* fix relative import

* fix os.makedirs exist_ok param compatibility PY2

* add compile flags in test_jit_load

* fix test_datasets unittest

* fix trt plugin clone and initialize bugs

* fix unit test error

* enable trt in ci py3

* update unittest timeout

* update lite tag

* disable ut

* Add cache for Transformer encoder.

* Bug fixed.

* add unittests for transformer encoder.

* dump to cpu

* format

* format

* format

fix push sparse int64 bug

* fix cc version

* fix

* add readme in whl package

* fix malloc L3 failed bug for kunlun

* minor