* "add model format design doc"

* "add restore function"

* "add parse protobuf"

* "move necessary information to saver.proto"

* "format code"

* "add gpu option"

* "add lod info"

* "add saveop python test wrapper"

* "checkpoint reuse save operator"

* "rewrite model format design doc"

* "async support needed"

* "fix run once"

* "fix doc based on comments"

* "refine based on comments"

* "fix based comments"

* "remove persistable flag from framework.proto"

* "add IndicateDataType to restore op"

* "add save test"

* "modify save restore code"

* "modified the restore logic"

* rm checkpoint_op.cc

* rm test_checkpoint_op.py

* "get inputs outputs name from execution context"

* Saving each variable to a independent file

* Fix bugs

* Rewrite save_restore_op_test with new Python framework

* Move `SaveOp` and `RestoreOp` from OpWithKernel to OpBase

* Refine unit test of SaveOp and RestoreOp

* fix compile errorwq

* Adding Adam Python wrapper

* Adding tests for Python Adam wrapper

survey on computation graph

 Enable to output LoD in fetch_op and check output LoD in the op unit test.

fix im2col interface

Add a warning to the docstring of Parameters.to_tar()

fix nccl compiler error

Add up, down, left and right padding for im2col.

deconv op implementing ...

* Add print_operators_doc.cc

* Update Escape

* Correct a bug

* Remove OpInfoMap::Iterate

* Update the print_operators_doc.cc

* Escape tab

* Use auto&

* Use auto&

* Remove trailing ,

* clang-format C++

Pass all forward op test

* Correct mul_op implementation

* Restore the origin shape after mul

* Fix mul op

* Do not touch math_function

* "add nccl enforce"

* Dev

* Update comment

* Add nccl test

* Follow comments

* init

* ensure ids in lookup table op must be a column vector

* add book4 configuration in test_layers

* debug test_book4

* add test_word2vec

* follow comments

* follow comments

* ensure ids in lookup table op must be a column vector

* follow comments

Fix a bug of dot-product attention.

Forward implementation for LSTM operator.

Fix v2 optimizer define order

fix crash on ds2

rename sparse_vector to sparse_float_vector, and fix typo.