* Fix batch norm if not track_running_stats

* test(BatchNorm2d): add autotest (#5734)

* Fix

Co-authored-by: Yinggang Wang <wyg19970408@gmail.com>

* rename rpc_token to TransportToken

* minor fix

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Refactor functional matmul and add apis.

* Export batch matmul and fix python module

* Check inplace valid in C++

* Support scalar add tensor.

* Support inplace when broadcasting add

* Refactor functional sub.

* Refactor mul and div

* Refactor functional sub.

* Fix div

* Fix add

* Fix add

* refacotr_maximum_minimum__py2cpp

* refine

* refine

* minor fix

* refactor

* auto format by CI

Co-authored-by: hjchen2 <chenhoujiangcug@gmail.com>
Co-authored-by: Yao Chi <later@usopp.net>
Co-authored-by: Yinggang Wang <wyg19970408@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* comm_net_sequence_number

* remove piece_id

* Remove IsAllowedActor

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* beautify stack frame

* make stack frame python style

* display more word

* refine

* add error type info

* add error type name

* refine and add summary_msg, msg info

* refine

* delete useless

* delete useless

* refine

* refine

* refine

Co-authored-by: Li Xinqi <lixinqi2010@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* pass test on linear with training

* Refactor RuntimeCtx for multi-runtime

* refactor inplace to support nn graph

* block support iterator

* block iter add check

* fix scalar_mul op conf build

* deal with inplace after merge master

* add alexnet graph test

* add cpu test and format

* cout to glog

* deal with Job run finish bug

* refactor lazy deal with inplace

* deal with 0D tensor

* update data path

* address review

* deal with lazy default attr

* mv according to ci

* merge master

* fix for ci

* fix for ci limit

* block proxy func

* support module custom func and refacotr get attr of block

* auto format by CI

Co-authored-by: chengtbf <472491134@qq.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* Add env ONEFLOW_THREAD_LOCAL_MESSAGE_QUEUE_ENABLE

* refine GetGlobalWorkStreamId

* refine name

* refine

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Refactor functional matmul and add apis.

* Export batch matmul and fix python module

* Check inplace valid in C++

* Support scalar add tensor.

* Support inplace when broadcasting add

* Refactor mul and div

* Refactor functional sub.

* Fix div

* Fix add

* Fix add

Co-authored-by: Yao Chi <later@usopp.net>
Co-authored-by: Yinggang Wang <wyg19970408@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* LazyInterpret::ApplyImpl for SourceUserOp and Copy

* Fix bug of CropMirrorNormalize ops; output ops check

* add test

* Add test case for tensor.to in nn.Graph

* remove useless code and refine error catch

* refine note

* Support CoinFlip module with consistent and device

* using const&

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Dev logical_and/or modules

* Add docs

* Add autotest

* Add single_client api

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* consistent_tensor_to_api

* add check_meta_consistency

* Update to.py

* refine

* refine

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* add avgpool 2d module

* pooling cpp

* make format

* fix format

* add avgpool1d but backward still have bug

* fix avgpool1d backward

* add avgpool3d cpu

* add avgpool series GPU kernel

* fix docs

* fix output grad size

* fix path

* fix comment

* fix comment

* fix comment

* remove get sequence

* fix doctest

* remove doctest part

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* add changes for multi dev demo

Signed-off-by: daquexian <daquexian566@gmail.com>

* add part of backward hook

Signed-off-by: daquexian <daquexian566@gmail.com>

* update

Signed-off-by: daquexian <daquexian566@gmail.com>

* add naive init_with_env

Signed-off-by: daquexian <daquexian566@gmail.com>

* update

Signed-off-by: daquexian <daquexian566@gmail.com>

* update

Signed-off-by: daquexian <daquexian566@gmail.com>

* support_multi_client

* update

Signed-off-by: daquexian <daquexian566@gmail.com>

* Remove unused code

* Fix multi client launch

* fix __main__ bug

* update abcd op

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix multi client sync, make nccl instr ordered

Signed-off-by: daquexian <daquexian566@gmail.com>

* temp changes

Signed-off-by: daquexian <daquexian566@gmail.com>

* Use functional api instead of op_expr_helper::XXXOp.

* align with latest master, remove unused code

Signed-off-by: daquexian <daquexian566@gmail.com>

* local rank returns 0 when no env var, save is_multi_client in EnvDesc

Signed-off-by: daquexian <daquexian566@gmail.com>

* move is_multi_client to ProcessCtx, rename cuda_d2d device to nccl, remove unused code

Signed-off-by: daquexian <daquexian566@gmail.com>

* abcd -> return_first_input op

Signed-off-by: daquexian <daquexian566@gmail.com>

* remove launch.py for now

Signed-off-by: daquexian <daquexian566@gmail.com>

* refine

Signed-off-by: daquexian <daquexian566@gmail.com>

* update IsMultiClient in env_util.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* rm multi_dev_demo.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* remove exported functions in env_util.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* remove unused op expr helper func

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix bug

Signed-off-by: daquexian <daquexian566@gmail.com>

* add DevVmDepObjectConsumeMode and set it as NONE in backward

Signed-off-by: daquexian <daquexian566@gmail.com>

* move return_first_input op from math_ops.py to tensor_ops.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix compile error

Signed-off-by: daquexian <daquexian566@gmail.com>

* refine

Signed-off-by: daquexian <daquexian566@gmail.com>

* add comments

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix exit bug in init.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* align with master

Signed-off-by: daquexian <daquexian566@gmail.com>

* update device ctor

Signed-off-by: daquexian <daquexian566@gmail.com>

* default dev id = local rank % gpu num

Signed-off-by: daquexian <daquexian566@gmail.com>

* assert single machine

Signed-off-by: daquexian <daquexian566@gmail.com>

* reformat

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix consume mode, implement eager_nccl_allreduce by process ranks

Signed-off-by: daquexian <daquexian566@gmail.com>

* fill sorted_ranks field in old code, reformat

Signed-off-by: daquexian <daquexian566@gmail.com>

* set default val for op conf, align with master

Signed-off-by: daquexian <daquexian566@gmail.com>

* impl return_first_input as functional api, impl allreduce as module

Signed-off-by: daquexian <daquexian566@gmail.com>

* add more tests

Signed-off-by: daquexian <daquexian566@gmail.com>

* reformat

Signed-off-by: daquexian <daquexian566@gmail.com>

* align with master

Signed-off-by: daquexian <daquexian566@gmail.com>

* rename ddp to flow.nn.parallel.DistributedDataParallel

Signed-off-by: daquexian <daquexian566@gmail.com>

* refine eager nccl comm

Signed-off-by: daquexian <daquexian566@gmail.com>

* refine eager nccl comm, divide grad by group size

Signed-off-by: daquexian <daquexian566@gmail.com>

* rename reversed_param_list -> ddp_state_for_reversed_params

Signed-off-by: daquexian <daquexian566@gmail.com>

* make return_first_input inplace

Signed-off-by: daquexian <daquexian566@gmail.com>

* restore eager allreduce

Signed-off-by: daquexian <daquexian566@gmail.com>

* add static all zero tensor and select first

Signed-off-by: daquexian <daquexian566@gmail.com>

* refine

Signed-off-by: daquexian <daquexian566@gmail.com>

* add functional allreduce op and use current rank group

Signed-off-by: daquexian <daquexian566@gmail.com>

* meterialize StaticAllZeroTensor in allreduce, support it in scalar mul

Signed-off-by: daquexian <daquexian566@gmail.com>

* materialize static zeros tensor in set_acc_grad

Signed-off-by: daquexian <daquexian566@gmail.com>

* rename

Signed-off-by: daquexian <daquexian566@gmail.com>

* auto format by CI

Co-authored-by: clackhan <han_binbin@163.com>
Co-authored-by: hjchen2 <chenhoujiangcug@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* pass test on linear with training

* Refactor RuntimeCtx for multi-runtime

* refactor inplace to support nn graph

* block support iterator

* block iter add check

* fix scalar_mul op conf build

* deal with inplace after merge master

* add alexnet graph test

* add cpu test and format

* cout to glog

* deal with Job run finish bug

* refactor lazy deal with inplace

* deal with 0D tensor

* update data path

* address review

* deal with lazy default attr

* mv according to ci

* merge master

* fix for ci

* fix for ci limit

Co-authored-by: chengtbf <472491134@qq.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Refactor functional matmul and add apis.

* Export batch matmul and fix python module

* Check inplace valid in C++

* Support scalar add tensor.

* Support inplace when broadcasting add

* Fix add

Co-authored-by: Yao Chi <later@usopp.net>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* debug

* construct job plan name as key

* rm debug code

* mod comment

* chang plan name to pubsub format

* correct comment syntax

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Rename CpuStreamType to AsyncCpuStreamType

* Rename CpuStreamType to AsyncCpuStreamType

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* fix_circular_deps_of_sbp_py

* address reveiw

* Revert "address reveiw"

This reverts commit 004ee6f89d7cec2972c77b40dc2659bd19078e2c.

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* fix

* fix

* address review

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* support tuple of ints size, or unpacked ints size for ones and zeros op

* refine according to comments

* add annotation for size

* support 0-dim tensor shape

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* export datasets interface

* auto format by CI

* fix docs

* add fix_datasets_export

* refine

Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>
Co-authored-by: tsai <jackalcooper@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Use function interface

* Delete the old interface file

* Delete the old interface file

* Delete the fifth part of the interface file

* Delete comment

* add RecordBytesDecoder

* add docs for OFRecordBytesDecoder

* update

* reformat the code

* resovlved conflits

* resolve conflicts

* resolving conflicts

* format the code

* Delete randperm.py

* format

* auto format by CI

* Update test_dataset.py

* auto format by CI

* The path in doctest fixed?

* Update test_dataset.py

* auto format by CI

* refine doctest

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* support_tensor_to/to_local

* export consistent_tensor.to_local()

* refine code

* export tensor.to()...

* refine code

* refine code

* optimize code

* refine code

* refine

* back up

* add tensor.to func

* make of_format

* remove to in pyTensor

* sync gpu data

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* backup

* refine

* rebase

* check in gen py

* merge master and fix bugs

* address pr comments

* address pr comments

* auto format by CI

* remove boxing

* refine

* Fix optional

* remove to in tensor.cpp

* update

* Support symbol placement type in functional.

* add sbp and sbp list arg

* refine

* use functional

* refactor CastConsistentOpExpr

* to_consistent(flow.B) backward

* Cache op expr

* add EagerNcclOpKernelState

* refine

* refine

* refine

* refine

* refine

* refine

* minor fix

* capture OpInterpContext

* unimplemented apply

* add GetNdSbp

* add mutex

* refine

* merge EagerConsistentTensorImpl::NewWithPhyTensor and EagerConsistentTensorImpl::NewWithoutPhyTensor into EagerConsistentTensorImpl::New

* rename functiona SyncData to SyncMetaAndData

* of_format

* add to_local to pybind

* add placement_sbp_util

* minor fix

* sync shape and data when tensor_to_local

* fix rpc_token bugs

* refactor AsyncRpcCtx

* set logical_shape correctly

* simplify implementation of consistent_tensor.to_local

* initialize rpc_token with zero

* refactor grad functions of to_consistent/to_local

* reformat and address pr comment

* reformat

* refactor eager_nccl_reduce lernel

Co-authored-by: tsai <jackalcooper@gmail.com>
Co-authored-by: Xinqi Li <lixinqi0703106@163.com>
Co-authored-by: Li Xinqi <lixinqi2010@gmail.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>
Co-authored-by: hjchen2 <chenhoujiangcug@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Use function interface

* Delete the old interface file

* Delete the old interface file

* Delet docs ctc_greedy_decoder

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* remove experimental

* add test num

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* add multi client multi machine test

Signed-off-by: daquexian <daquexian566@gmail.com>

* remove copy cores

Signed-off-by: daquexian <daquexian566@gmail.com>

* use discover in bash

Signed-off-by: daquexian <daquexian566@gmail.com>

* add tests in test.yml and refine

Signed-off-by: daquexian <daquexian566@gmail.com>

* remove multi_client test files into test dir to reuse code

Signed-off-by: daquexian <daquexian566@gmail.com>

* delete distributed_run_multi_client.py and move impl in distributed_run.py

Signed-off-by: daquexian <daquexian566@gmail.com>

* if -> elif

Signed-off-by: daquexian <daquexian566@gmail.com>

* try three times and upload log

Signed-off-by: daquexian <daquexian566@gmail.com>

* add 'mode' arg in py

Signed-off-by: daquexian <daquexian566@gmail.com>

* auto format by CI

* remove --multi_client in yml

Signed-off-by: daquexian <daquexian566@gmail.com>

* skip distributed test in cpu

Signed-off-by: daquexian <daquexian566@gmail.com>

* use new test container

Signed-off-by: daquexian <daquexian566@gmail.com>

* add host key to all machines

Signed-off-by: daquexian <daquexian566@gmail.com>

* auto format by CI

* fix python version

Signed-off-by: daquexian <daquexian566@gmail.com>

* fix python version

Signed-off-by: daquexian <daquexian566@gmail.com>

Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Multi-Runtime for multi nn.Graph

* fix import os

* Fix BUG of IsMultiClient

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* 0-dim tensor support

* test case

* add more test

* refine

* update

* update default constructor

* reconstuct

* merge master

* remove notes

* remove useless codes

* fix comments

* fix comment

* add test case

* format

* refine

* refine

* refine

* refine

* MirroredTensorMeta::MirroredTensorMeta()

* support 0-dim slice

* support 0-dim slice grad

* refine

* auto format by CI

* refine

* refine

* auto format by CI

* refine

* fix slice bug

* auto format by CI

* fix resnet50 0-im loss uasge

* fix 0-dim tensor usage in test cases

* add skip test

* auto format by CI

* fix test_dataset

* check blobdesc.shape init

* auto format by CI

* remove useless empty shape init

* fix l1loss 0-dim error

* auto format by CI

* fix argmax op test

* fix add_n op test

* auto format by CI

* fix bce loss op test

* auto format by CI

* fix squeeze op test

* fix conv2d op test

* fix xpu_shape for clip_grad_norm

* auto format by CI

* resolve confilct

* fix multi-cpu slice_copier 0-dim bug

* auto format by CI

* add memory copy for 0-dim

* auto format by CI

* support copy0dim

* refine

* auto format by CI

* remove unuse codes

* fix check for kldivloss

* gpu 0-dim copy

* auto format by CI

* fix clip_grad_norm doctest

* fix reduce_ops doctest

* fix argmax doctest

* fix loss module doctests

* fix math_ops doctests

* fix norm modules doctest

Co-authored-by: Xinqi Li <lixinqi0703106@163.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* add conv3d module

* add simple test case

* use conv base class to write conv3d

* still test error

* add torch style conv3d unit test

* fix format

* add assert

* unittest still error

* auto format by CI

* fix format and autotest

* remove dir

* remove useless file

* add extra expr

* auto format by CI

* fix import

* fix doc

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Fix optional

* update

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* add batchnorm3d module

* add testing flag, add more log

Signed-off-by: daquexian <daquexian566@gmail.com>

* only test cpu device

* auto format by CI

Co-authored-by: daquexian <daquexian566@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* feat(Tensor): support 0shape tensor

* math binary broadcast support emoty tensor input

* slice support empty tensor input and output

* fix check in slice

* test(Cat): add 0shape cat module test

* fix return type error on gcc 4.8.5

Signed-off-by: daquexian <daquexian566@gmail.com>

* auto format by CI

* add module op test for empty tensor, cuda kernel support empty tensor

* format

* feat(ReduceOp): reduce op kernels support 0shape tensor

* delete files added by mistake

* refine if

* refine if

* feat(ConstantOp): constant ops support 0shape tensor

* feat(ReshapeOp): reshape kernel support 0shape tensor

* math binary and unary backward skip when elem equal to zeros

* fix(ReduceOp): fix reduce not memset bug

* support getitem output empty tensor

* fix comment

* getitem support input is empty

* reduce_like kernel support empty

* fix op test bug

* feat(ReduceOp): refine reduce ops initialize value

* format code

* fix triu bug when input is empty

* test(AbsOp): fix test bug

* test(DivOp): fix test bug

* fix clamp bug

* fix test_sub bug

* fix(ReduceOp): fix reduce op memset bug

* auto format by CI

* fix random

Co-authored-by: liufengwei <2472937968@qq.com>
Co-authored-by: daquexian <daquexian566@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

* fix bug

* add cast_like functor

* convert to functional impl

* convert to functional

Co-authored-by: Houjiang Chen <chenhoujiangcug@gmail.com>
Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>

* Move Global<CommNet> to env scope

* Revert Global RuntimeCtx New

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>
Co-authored-by: cheng cheng <472491134@qq.com>

* rewrite slice_update backward

* use functional api

* fix comments

* refine

* refine

* refine

* add slice_grad functor

* refine

* auto format by CI

Co-authored-by: oneflow-ci-bot <ci-bot@oneflow.org>

Co-authored-by: oneflow-ci-bot <69100618+oneflow-ci-bot@users.noreply.github.com>