* fix dygraph double grad, test=develop

* fix unpack constructor, test=develop

* sequential reader stage 1, test=develop

* fix ut, test=develop

* fix iterable=False reset bug, add some logs and polish code, test=develop

* inference feed partial data, test=develop

* Turn on keep_order=True for test, test=develop

* enhance ut to test more cases, test=develop

* test commit for reverting

* Revert "test commit for reverting", test=develop

This reverts commit 80aef42e.

* add ut of merged and unmerged results, test=develop

* add more uts for coverages and add en doc of api, test=develop

* follow comments, test=develop

* change note style, test=develop

update embedding_eltwise_layernorm fuse pass and fused kernel, to support multi input

* add double grad implementation for dygraph, test=develop

* polish code, add uts, test=develop

* fix place bug, test=develop

* polish codes, add more uts for coverages, test=develop

* add no_grad_set, test=develop

* add star gan ut, test=develop

* follow comments, test=develop

As the title

* move __getitem__ to cpp

* bug fix

* add type check and gil release

* support negative step with omitted ends
test=develop

* code refine
test=develop

* bug fix
test=develop

* slice always return different pyobj
test=develop

* Support to use external function test=develop

* refine the parms of ast_to_func test=develop

* support to modify dict and list in while_loop test=develop

* polish code test=develop

test=develop

* fix div zero test=develop

* fix div zero test=develop

* add hostdevice function test=develop

* add eps when is zero test=develop

1. Add support for Break and Continue in Dygraph to Static
2. Also add support for gast.Not in NodeTestTransformer
3. Also add support for logical op transformation in LoopTransformer

initialize global nccl context in dygraph
test=develop

* don't add unregisterd attr, test=develop

* add some unittests,test=develop

This reverts commit 66fce9e8.

* remove unnecessary prepare data, test=develop

* Op in while block will not skip PrepareData, test=develop

* Support and/or in controlFlow if test=develop

add approximation for gelu, default value is False (only kernel with eigen is added, remove code for computing gelu with MKLDNN temporarily)

test=develop

* test=cpu

* test=cpu

* test=py2gcc8

* test=py2gcc8

* test=py2gcc8

* test=py2gcc8

* test=py2gcc8

* test=py2gcc8

* test=py2gcc8

* test=develop;test=document_fix;test=py2gcc8

Fleet Parameter Server API Integrated

* Improved stability through the use of caching

test=develop

* fix style

test=develop

* Fix syntax error

test=develop

* Fix syntax error

test=develop

* Support Tensor.shape in control flow if/for/while and separate TensorShapeTransformer from BasicApiTransformer. test=develop