* [Auto Parallel] Replace the old planner by the new partition tuner

* [Auto Parallel] Improve the completion and distributed context

* [Auto Parallel] Fix some bugs of the compatible check of some dist ops

* [Auto Parallel] Fix some bugs

* add cuda11.4 develop docker

* change default python from 2.7 to 3.7

* change base image for cpu docker

* fix gcc bug

* fix whl package name

* update go version and delete useless package in dockerfile

* fix release18 error

* fix wget sqlite problem

* update go version

* update go version in dev dockerfile

* fix CI error

* install zstd

* fix CI error

* add --no-check-certificate when install go

* python2.7 do not install requirements

* fix CI Coverage error

* coverage==5.5

* fix test_activation ut

* let numpy < 1.22 to pass test_activation_op unittest

* fix test_python_bf16_numpy_datatype unittest

* change paddle-bfloat==0.1.3

* recover version of paddle-bfloat

* bind elementwise_mod_op_xpu *test=kunlun

* add more supported dtypes and UTs *test=kunlun

* fix datatype error

* add op to in xpu1_op_list

* Update Mac cmake version >=3.15 (#41456)

* Update Mac cmake version >=3.15

* notest;read test1

notest;read test2

notest;read test3

* fix inference link error

* fix inference link error

* fix windows link error

* fix cmake_policy

* fix build big size

* Add paddle::variant and replace paddle::any (#42139)

* add variant and replace any

* split attribute

* disable unittest failed in eager CI in temporary (#42101)

* test=py3-eager

* test=py3-eager

* test=py3-eager

* combine graph_table and feature_table in graph_engine (#42134)

* extract sub-graph

* graph-engine merging

* fix

* fix

* fix heter-ps config

* test performance

* test performance

* test performance

* test

* test

* update bfs

* change cmake

* test

* test gpu speed

* gpu_graph_engine optimization

* add dsm sample method

* add graph_neighbor_sample_v2

* Add graph_neighbor_sample_v2

* fix for loop

* add cpu sample interface

* fix kernel judgement

* add ssd layer to graph_engine

* fix allocation

* fix syntax error

* fix syntax error

* fix pscore class

* fix

* change index settings

* recover test

* recover test

* fix spelling

* recover

* fix

* move cudamemcpy after cuda stream sync

* fix linking problem

* remove comment

* add cpu test

* test

* add cpu test

* change comment

* combine feature table and graph table

* test

* test

* pybind

* test

* test

* test

* test

* pybind

* pybind

* fix cmake

* pybind

* fix

* fix

* add pybind

* add pybind
Co-authored-by: NDesmonDay <908660116@qq.com>

* [CustomDevice] add eager mode support (#42034)

* fix FlattenContiguousRangeOpConverter out dim error (#42087)

* fix FlattenContiguousRangeOpConverter out dim error

* update code

* fix python3.10 compile bug on windows (#42140)

* Optimize dygraph GetExpectedKernelType perf (#42154)

* opt dygraph scheduling

* revert part impl

* fix incorrect usages of std::move and other compile errors (#41045)

* fix bug of std::move and others

* fix an compile error in debug mode

* fix wrong copy assignment operator
Signed-off-by: Ntiancaishaonvjituizi <452565578@qq.com>

* reformat
Signed-off-by: Ntiancaishaonvjituizi <452565578@qq.com>

* reformat
Signed-off-by: Ntiancaishaonvjituizi <452565578@qq.com>

* fix ArrayRef constructor following llvm

* fix format

* fix conflict with master

* fix variant compile error (#42203)

* [Eager] Support numpy.ndarry in CastNumpy2Scalar (#42136)

* [Eager] Remove redundancy code, fix fp16 case (#42169)

* [Eager] Support div(scalar) in eager mode (#42148)

* [Eager] Support div scalar in eager mode

* Updated and remove debug logs

* Remove list, use 'or' directly

* Remove useless statement

* fix recompute (#42128)

* fix recompute

* modify return

* add LICENSE in wheel dist-info package (#42187)

* replace any by variant in infermeta (#42181)

* 【PaddlePaddle Hackathon 2】24、为 Paddle 新增 nn.ChannelShuffle 组网 API (#40743)

* Add infermeta for ChannelShuffle

* Create channel_shuffle_grad_kernel.h

* Create channel_shuffle_kernel.h

* Create channel_shuffle_sig.cc

* Create channel_shuffle_op.cc

ChannelShuffle算子的描述

* Create channel_shuffle_kernel_impl.h

ChannelShuffle核函数的实现

* Create channel_shuffle_grad_kernel_impl.h

ChannelShuffle反向核函数的实现

* Add kernel register of channel shuffle and grad

注册ChannelShuffle及其反向的核函数

* add nn.functional.channel_shuffle

* add nn.ChannelShuffle

* Create test_channel_shuffle.py

* Update example of ChannelShuffle in vision.py

* Update test_channel_shuffle.py

* 修改channel_shuffle核函数的实现位置

* 修正代码格式

* 删除多余空格

* 完善channel_shuffle的错误检查

* Update unary.cc

* Update channel_shuffle_op.cc

* Update test_channel_shuffle.py

* Update unary.cc

* add channel_shuffle

* Update test_channel_shuffle.py

* Update vision.py

* 调整代码格式

* Update channel_shuffle_sig.cc

* 更新ChannelShuffle的文档

* 更新channel_shuffle的文档

* remove ChannelShuffleOpArgumentMapping

* add ChannelShuffleGradInferMeta

* Update channel_shuffle_op.cc

* 调整channel_shuffle及其梯度的核函数的位置

* Do not reset default stream for StreamSafeCUDAAllocator (#42149)

* remove redundant computation in Categorical.probs (#42114)

* Downloading data for test_analyzer_vit_ocr (#42041)

* Change server URL

* update config

* add test to parallel UT rule

* add checksum to ensure files are downloaded

* change downloading target

* reuse existing variable

* change target directory

* fix en docs of some Apis (gradients, scope_guard, cuda_places, name_scope, device_guard, load_program_state, scale, ParamAttr and WeightNormParamAttr) (#41604)

* Update scope_guard; test=document_fix

* gradients; test=document_fix

* gradients; test=document_fix

* name_scope; test=document_fix

* cpu_places; test=document_fix

* WeightNormParamAttr; test=document_fix

* cuda_places; test=document_fix

* load_program_state; test=document_fix

* device_guard; test=document_fix

* device_guard; test=document_fix

* ParamAttr; test=document_fix

* scale; test=document_fix

* scale; test=document_fix

* update code example；test=document_fix
Co-authored-by: NChen Long <1300851984@qq.com>

* fix datatype error

add op to in xpu1_op_list

*test=kunlun

* fix elementwise_mod op path error  *test=kunlun

* fix elementwise_mod UT error  *test=kunlun

* fix datatype error

add op to in xpu1_op_list

*test=kunlun

add op to in xpu1_op_list

fix elementwise_mod op path error  *test=kunlun

fix elementwise_mod UT error  *test=kunlun
Co-authored-by: Ntianshuo78520a <707759223@qq.com>
Co-authored-by: NChen Weihang <chenweihang@baidu.com>
Co-authored-by: Npangyoki <pangyoki@126.com>
Co-authored-by: Nseemingwang <seemingwang@users.noreply.github.com>
Co-authored-by: NDesmonDay <908660116@qq.com>
Co-authored-by: Nronnywang <524019753@qq.com>
Co-authored-by: Nbaoachun <962571062@qq.com>
Co-authored-by: Zhou Wei <1183042833@qq.com>
Co-authored-by: Ntiancaishaonvjituizi <452565578@qq.com>
Co-authored-by: NWeilong Wu <veyron_wu@163.com>
Co-authored-by: NRoc <30228238+sljlp@users.noreply.github.com>
Co-authored-by: NBrilliantYuKaimin <91609464+BrilliantYuKaimin@users.noreply.github.com>
Co-authored-by: NRuibiao Chen <chenruibiao@baidu.com>
Co-authored-by: NFeiyu Chan <chenfeiyu@baidu.com>
Co-authored-by: NSławomir Siwek <slawomir.siwek@intel.com>
Co-authored-by: NYilingyelu <103369238+Yilingyelu@users.noreply.github.com>
Co-authored-by: NChen Long <1300851984@qq.com>

* [NPU] add clip_by_norm op

* fix

* update

* fix dataset ut

* fix seed state ut

* rm transfer_cast_op_pass

* rm header

* update UTs 0

* fix ci

* fix ci 3

* add default_ctx in backward.py

* record grad_var_to_var with grad_times

* fix backward

* update annotation

* add complete_high_order_grad in complete_forward

* add dist slice op

* update grad_var_to_var type

* update partition_block init mapping before loss op

* update compatible for 'XShape' & update 'allreduce_vars'

* add dist reshape op when input dim equal to output dim

* update 'set_grad_var_shape' with grad_var_to_var

* fix dist slice

* fix set_grad_var_shape

* add dist pnorm op

* fix dist pnorm dist_attr

* fix engine startprogram & adapt highorder grad

* fix set_grad_var_shape when mp

* update unittest

* update cmakelist

* default strategy in engine: dp

* bug fix

* tiny fix

* flatten outputs

* fix default strategy

* init default ctx

* tiny fix

* test=allcase

* merge recent changes

* fix setting pipline

[New API] add API paddle.vision.transforms.RandomErasing and paddle.vision.transforms.erase (#42280)

* add api RandomErasing
* add numpy/PIL backend support and UT
* fix doc and optimize UT
* add seed

This PR hotfixed the `test_cond.py` in CUDA 11.2

The reason of the bug is that the `fill_constant` op returns wrong value in the modified test case `test_extremely_simple_net_with_op_in_condition`, SWEs can use `layers.Print(a)` and `layers.Print(b)` in the test case to reproduce it and they can see the `fill_constant` returns something `e-50` instead of `1.23` and `1.25`

This PR hotfixed the bug by comparing `b` value instead of actual number, which makes sure the `cond` logic is right. **However, the PR didn't fix `fill_constant`**. We would let the SWEs who are working here to find the op bug and fix it.

* add Tensor support for sub-functions of colorjitter
* add UT

* Extend python einsum interface to make einsum_v2 support multi-operands and switch it to default.

* add opt_einsum dependence

* add yaml and support eager model

* fix by code review

* [Eager] Remove enable_legacy_dygraph setting

* Add more tests

* add double yaml

* add inline func

* [Dy2Stat]Fix losting pre/post hook from outermost layer while jit.save

* fix kwargs

* fix unittest

* add gradient merge for DistributedFusedLamb

* use master acc gradient

* fix CI ut

* polish

* remove math_function_impl.h change

* fix test_update_loss_scaling_op.py

* try to fix XPU/NPU CI

* add gm ut

* fix PIL sample mode deprecated warning

* compatible with old pil version