* only run pre-commit

* only run pre-commit

* add unbalanced data

* fix utest

* reformat hybrid_parallel_util.py by black

* add fused_allreduce_gradients_with_group

* add scale

* fix ci

* Fix bug of reduce_sum op. When input.numel() > INT32_MAX, its result
is wrong.

* support pure bfloat16

* support bf16 linear

* update PR to pass CI

* tiny fix where_grad_kernel.cu

* Support bfloat16 type for reducer and sharding.

* Fix some bug.

* Polish code.

* Polise code.

* Add bfloat16 datatype in fill_grad kernels.
Co-authored-by: Nsneaxiy <sneaxiy@126.com>
Co-authored-by: Nsneaxiy <sneaxiy@126.com>

* Fix bug of reduce_sum op. When input.numel() > INT32_MAX, its result
is wrong.

* Support allow_partial switch, which can be configure in
pipeline_configs. If sent tensor are not the same from
different hosts, they shouldn't been sent partially and
then concated as a whole tensor.

* Change name allow_partial to enable_partial_send_recv.

* Add global variable _enable_partial_send_recv

* [dygraph sharding] Overlap the reduce and the caculation for sharding stage 2. (#46495)

* [dygraph sharding stage 2] sharding broadcast overlap (#46656)

* Multi groups for broadcast of sharding stage 2 (#46894)

* [Dygraph] Fix performance of pp+mp by using send/recv_calc_stream instead of send/recv (#46116)

* [Dygraph] Fix Perf of FusedFeedForward and FusedAttention with AllReduce (#46780)

* update

* Support both use_calc_stream and sync_op in send recv APIs (#46023)

* Support both use_calc_stream and sync_op in allgather API (#46295)

* Support both use_calc_stream and sync_op in collective communication API (#46761)

* Move group and all reduce from collective to communication (#45848)

* Completes bfloat16 dtype for collective api in eager mode (#45844)

* Fix collective APIs cannot be recognized when building docs (#46962)
Co-authored-by: NLiYuRio <63526175+LiYuRio@users.noreply.github.com>

* bug fix for virtual pipeline parallel (#45922)

* dont wait for send op under dygraph pp (#46209)

* [interleave pp] sync recv for 1f1b (#46399)

* [dygraph pp] all sync for allgather partial (#46483)

uniform logger manager in FleetAPI.
hidde API under distributed/utils which users don't need.

* fix bugs of mp

* fix bugs of mp

* update

* update

* fix bug

* polish code comments

* polish data_device_transform.cc

* [Auto Parallel] Change the import way of Auto Parallel (#46115)

* fix strategy (#46256)

* [Auto Parallel] performance improvement for Sharding-DP hybrid parallelism (#46180)

* remove no need grad allreduce communication when sharding-dp

* remove no need grad allreduce communication when sharding-dp

* bugfix

* bugfix

* bugfix
Co-authored-by: NYulong Ao <aoyulong@baidu.com>
Co-authored-by: NJZ-LIANG <jianzhongliang10@gmail.com>

* refactor mp.

* update setup.py.

* update mp_layers.py for compatibility.

* add documents for mp_layers.py

* update init.py

* update collective.py.

* update.

* update mp_ops.py

* update.

* update code style.

* update code style.

* support iterable dataset for auto parallel

* add split_data proto

* fix unittest bug

* fix recompute bug

* update cmake

* config

* fix unittest

* zero init & cache & patch config

* add barrier to save and load

* add unittest

* add quant pass

* reconstruct_of_fleet_api

* update

* refactor fleet.

* refact fleet.py.

* update fleet/__init__.py.

* update fleet.py

* update code style.

* update fleet

* update fleet

* update fleet

* update fleet

* update model.py

* update fleet.

* update __init__.py

* update fleet.

* update fleet.

* update fleet

* update fleet

* update fleet

* update fleet.

* update optimizer.py

* update optimizer

* update fleet.py

* update scaler.py

* update setup.py.in

* back fl

* delete ssl cert

* .

* make warning

* .

* unittest paral degree

* solve unittest

* heter & multi cloud commm ready

* .

* .

* fl-ps v1.0

* .

* support N + N mode

* .

* .

* .

* .

* delete print

* .

* .

* .

* .

* fix bug

* .

* .

* fl-ps with coordinator ready

* merge dev

* update message parse only

* update fl client scheduler

* fix bug

* update multithreads sync

* fix ci errors

* update role_maker.py

* update role_maker.py

* fix ci error: windows py import error

* fix ci error: windows py import error

* fix windows ci pylib import error

* add dump fields & params

* try to fix windows import fleet error

* fix ps FLAGS error

* fix logging risk

* fix logging possible risk

* write trainer_desc file

* support split sparse params in local & remote

* fix import paddle.fluid.core.PSGPU

* fix import paddle.fluid.core.PSGPU

* add remote_sparse & local_sparse config

* fix unittest

* fix test_dist_fleet_geo table error

* fix PADDLE_ENFORCE error

* fix other's pr conflict

* [OpAttr]Support VarDesc* and vector<VarDesc*> in Attribute

* add unittest for inference predictor

* save load

* save load

* add unittest

* first commit

* second commit

* third commit

* remove SaveLocalFS in memory sparse table

* save dense param

* update

* push slot

* fix push show clk: int -> float

* add unittest

* fix sample

* unittest

* add AsExtra for op

* unittest

* modify fs.py

* modify fs.py

* fix some bugs

* add dataset hdfs config

* local change

* dataset use differenct hadoop ugi/fs_name

* add

* fix conflict

* fix

* remove logs

* code style

* fix

* code style

* code style

* fix

* code style

* save_dense_param

* fix

* fix

* fix

* fix

* change momentum in dense optimzer

* fix

* fix

* change fluid => paddle.static

* remove some unuseful code
Co-authored-by: Nesythan <esythan@126.com>