paddle.distributed.fleet supports dynamic graph execution.

* support running python train.py for fleet-task; test=develop

* update set_dict method name & add aliases

* fix var name error

* fix alias formats

* use set_state_dict in unittest

* add decorator solve compatible problem

* polish decorator

* replace layer set_state_dict by patched method

* remove import monkey path layer

* fix import function error

* add unittest for coverage

* support load infer model format state dict

* add unittests

* remove keep name table

* recolve circle inport

* fix compatible problem

* recover unittest

* polish doc and comment

* set_default_type only take effect on python floats or complex

* fix doc

* refine paddle.stack

* support TensorArray

* add test

* fix coverage problem

* fix coverage problem

* fix sample code

* fix uniform CI random fail problem

* decrease diag_num in DiagInit

* fix dropout bug in backward when input is 1d tensor, test=develop

* add test case and refine error message, test=develop

* refine error message, test=develop

As the title

* fix norm bug, test=develop

* fix norm bug, test=develop

* fix norm bug, test=develop

* fix norm bug, test=develop

* fix norm bug, test=develop

Raise RuntimeError if run the callable object decorated by '@paddle.jit.to_static' not in dynamic mode.  (#26750)

* fix the argmin,argmax op for the paddlepaddle 2.0， add checkPoint for the argmax/argmin

* fix fluid.embedding

[Dy2Stat-debugging] Fix bug: Returns True if the verbosity level set by the user is greater than or equal to the log level (#26858)

Fix an English typo in unit test.

* support mnist and resnet dygraph_to_static test

* make FLAGS_use_mkldnn a public flag

* fix test_get_set_flags

* Change name of a function

* Rerun CIs commit

* Fix oneDNN dygraph training
Co-authored-by: NAdam <38704900+grygielski@users.noreply.github.com>
Co-authored-by: Ngrygielski <adam.grygielski@gmail.com>

* add embedding 2.0

* add embedding support input int32

* update doc

* update doc

* fix optimizer sample code

* add default value for adamw weight_decay

* fix adamw

* change LearningRateDecay to _LRScheduler

* fix adamw;notest

* fix load;notest

* remove file

* bug fix

* fix code style

* bug fix

* add ut

* adamw support weight_decay=0

* fix ut

* fix set_lr doc

* fix doc

* change parameters place

* remove backend argument of init_parallel_env

* remove keep name table in transformer

* add cpu version check

* add skip unittest for init_parallel_env

* polish doc: remove func use & update example

rename Dropout2D/3D to Dropout2d/3d

* refine cudnn softmax

* get use of global 'use_mkldnn' in layer_helper

* update for CI

* update for CI, relu test

* update for CI, relu test added, make FLAGS_use_mkldnn a public flag

* added more strict tests, fixes after review

* fixes after review

* fixes after review, CI stuff

* Fix Python Version compatibility

* add import six

* Move hapi form paddle/incubate to paddle

* Remove vision/datasets/utils.py and clean code

* Add sample code for conll05

* Print pull path when saving model

* Fix sample code after paramter_list of SGD is changed to parameters

* Fix bug in wmt16 datase

* set default dtype for distribution API

* Add unittest

* Add unittest

* fix import get_default_dtype problem

* delete change under fluid.layers.nn

* little change

* Support Heter Parameter Server