* immigrate_transpose_to_pten cpu kernel only; test=develop

* fix bug; test=develop

* add transpose cuda api

* bug fix;

* fix bugs

* fix bugs; test=develop

* bug fix;

* move transepose to pten; test=develop

* fix bug; test=develop

* fix bugs; test=develop

* add transpose grad fp16 support; test=develop

* fix bug; test=develop

* fix npu bug; test=develop

* fix nemul = 0 bug; test=develop

* add fp16 support; test=develop

* fix data type register bug; test=develop

* fix transpose bug; test=develop

* update transpose

* fix transpose bug; test=develop

* remove useless code; test=develop

* remove useless code; test=develop

* fix transpose alias bug; test=develop

* polish code; test=develop

* resolve confict; test=develop

* resolve confilct; test=develop

* recover prepared operator; test=develop

* fix bug; test=develop

* polish code; test=develop

* fix bug; test=develop

* fix bug; test=develop