* refine sum_op mkldnn code

* refine sum_op mkldnn code

* add fp16 data type for set_value

* cancel flip modification

* add fp16 dtype support for set_value

* solve transpose2, follow #22402

* fix CI cmake

* update REGISTER_OP_KERNEL of transpose2

* Update README.md

* Update README.md

* Fix E713: convert 'not ... in' to 'not in'

* Fix some bugs hidden in build_cinn_pass.

* Update codes about OpTransInfo.

* Only support for the static reshape/reshape2 op.

* Add libpaddle.so log

* Add libpaddle.so log

* Update README.md

* Update README.md

* Fix F632: replace 'is', 'is not' with ==, != respectively

* Update README.md

* Update README.md

* Update .flake8 config: remove F601 code, text=document_fix

* Update README.md

* Update README.md

* Fix F601: only keep the last key among the same names

* Update README.md

* Update README.md

* Fix F541 by converting f-string to normal strings

* Update README.md

* Update README.md

* Remove F402 code in config

* Update README.md

* Update README.md

* Fix F402 by changing the name of loop variable

* Change variable name
Change variable name 'temp_tensor' to 'var' to avoid confusion, as var
also appears in the same file.

[CodeStyle][F401] update flake8 F401 config (python_paddle/proto,onnx,optimizer,hapi,autograd,framework,io,cost_model,geometric,vision,static). test=document_fix (#46756)

[CodeStyle][F401] remove unused imports in python_paddle/io_costmodel_geometric_vision_static. (#46755)
Signed-off-by: Yulv-git <yulvchi@qq.com>
Signed-off-by: Yulv-git <yulvchi@qq.com>

[CodeStyle][F401] remove unused imports in python_paddle/proto_onnx_optimizer_hapi_autograd_framework. (#46754)

* add yaml entry for rnn and rrnn_grad, move infershape function for rnn_grad to phi infer meta

* WIP: move rnn kernrl to phi

* Change the code generation to avoid converting from intializer list to tuple of heterogeneous types.
This is only triggered when an api has intermediate outputs, and the result of the outputs are of heterogeneous types.

* fix the bug that when none in a vector of tensors requires gradient, the conversion to InferShapeContext to InferMetaContext (a.k.a. BuildInferMetaContext) produces errorous results.

* fix ci bugs

* fix ci bugs

* fix ci bugs

* modify code according comment
Co-authored-by: Nchenfeiyu <chenfeiyu@baidu.com>

* [CodeStyle][F401] remove unused imports in unittests/distributed_passes,tokenizer,sequence

* add noqa after required imports

* reduce time cost on atomic in interpretercore

* clear code of PrepareAtomic in interpretercore

* refine threadpool cache