* replace assert false with AssertionError

* 修改配置文件多余的部分

fix embd plugin: S = mask_id.d[1]

* fix batch_norm custom_vjp dtype

* add nhwc test example

* fix typo

* [Fix typo] Fix typo error in grad_node_info.h

* fix varbase_patch_methods.py typo error

* fix more typo errors

* remove fluid.data

* fix typro

* fix somme unitest error

* fix conflicts

* fix sample code error

* fxi sample coder error

* fxi sample code error

* fxi sample code error

* fix xpu test error

* fix xpu test error

* Delete ps_pb2.py

* fix test error

* fix typro

* fix sample code error

* fix comments

* fix test norm op data

* fix sample code error

* fix conflicts

* fix build_ci_problem

* fix_bug,test=document_fix

* init

* close cinn

* close cinn

* add public_python_api

* fix GN when params.c% params.cPerBlock != 0

* fix GN when params.cnot divisable by params.cPerBlock

* simplify batch_norm composite rule

* polish code

* add public_python_api in prim test

* fix test_squeeze2_op

remove namespace

codestyle

move setPlace to Public

fix devicetype

* move DataType from paddle::experimental to phi

* convert namespace

* convert namespace

* convert namespace

* clarify namespace

* convert more datatype

* Revert "convert more datatype"

This reverts commit 083b462959e6a22d4d8767707b628b95b396642e.

* convert more in auto_code_generator

* fix conflicts for XPU

* fix namespace conflicts

* fix errors

* Revert "fix errors"

This reverts commit f9d9958b54ee32141112274c8a5c3c381ab0f876.

* fix errors

* fix formatting

* fix flash_attention

* Update mp_layers.py

* set order for hybridparallel

* fix bug

* fix ->

* fix ->

* fix ->

* add topology

* fix utest

* add patterns

* add unittest

* [Zero-Dim] Support output 0D for argmin/argmax/median/kthvalue/mode/equal_all/allclose

* fix CI

* fix divide zero bug for softmax_with_cross_entropy

* change the single test way

* fix the null pointer in graph_khop_sampler

* fix the case5

* add fp16 unittest

* support bf16 and add unittest

* fix according to review

* support relue custom vjp

* Fix some typo issue for relu custom vjp

* [XPU] add fp16 support for compare ops.

* fix ci.

* with printf

* add DropOutNdForwardKernel

* PR comment

[Zero-Dim] Support 0D for numel/rank/size/optimizer/create_parameter/create_global_var, fix some usage to adapt 0D (#51566)

* fix add axis is not default -1

* polish conv logic

* Don't import paddle

* fix error