Add basic while_loop

Add tests to use dy/dx to make sure the gradient values calculated by the control flow backward is correct. Also fixed bugs detected by those tests.

Fix bugs:

1. Unlike sum_op, optimizer ops don't allow uninitialized input tensor. But in conditional_block_grad_op, since the conditional_block may not run, the output gradient tensor may be uninitialized, which will cause the optimizer op error. To fix it, we should let optimizer ops support uninitialized input like sum_op or assign the uninitialized gradient to 0 when the conditional_block_grad_op doesn't run. I found there are about 10+ optimizer ops. **To be simpler, I just assign output gradient of the conditional_block_grad_op to 0 in this PR**. But it can be further explored whether we can make optimizer ops like sum_op to support uninitialized input tensor because theoretically we can speed up without the assigning in conditional_block_grad_op.

2. Infer parameter shapes during append_backward. I didn't know that all our parameters are in global block. When op_desc is inferring shapes at the sub-block, it may not know the shape of gradients of parameters whose shape information is at global block. I fixed it by inferring shapes of gradients from forward var.

This PR also did some code clean up:
1. Print the var name when sgd_op catches shape error so that it is easier to debug
2. Fix a typo: dicta -> dict

Add English doc for cond

* Commit before merging develop

test=develop

* Backup after working with Huihuang logs

* Commit before deleting Huihuang debug loggings

* Commit before debug

test=develop

* Fix bug commit

test=develop

* Backup of fixing bugs

test=develop

* Clean up code

test=develop

* Fix a bug in sum_op

test=develop

* add input type and input data type check for Print_op test=develop

* add control flow API: case. test=develop

* delete 'raise TypeError' in _error_message() and return a string. test=develop

* polish API document. test=develop

* add API switch_case. test=develop

add Nest

* modify code according to reviews:
1.Attr(branch_index) support 'uint8' and 'int64' besides 'int32'.
2.remove useless code.
test=develop

* replace fluid.layers.data with fluid.data and polish API document. test=develop

These ops are useful in control flow.

* Refine create_array api en doc

test=develop
test=document_fix


* Fix api doc example bug and polish square

test=develop
test=document_fix


* Refine comment

test=develop
test=document_fix


* refine API.spec

test=develop
test=document_fix

refine the en api doc of ones, zeros, reverse, assign, increment, hsigmoid and create_py_reader_by_data ops (#20343)

* refine the en api doc of ones, zeros, reverse, increment, hsigmoid and create_py_reader_by_data ops
test=develop, test=document_preview, test=document_fix

* refine eng doc for hsigmoid and create_py_reader_by_data ops
test=develop, test=document_preview, test=document_fix

* update API.spec
test=document_fix

* Fix the parameter name axis of reverse op in eng doc
test=develop, test=document_fix

* Update API.spec
test=develop, test=document_fix

* Refine eng doc of zeros, ones, reverse and assign op
test=develop, test=document_fix

* Update API.spec for assign, ones, zeros and reverse
test=develop, test=document_fix

* Fix data type of reverse op in eng doc
test=develop, test=document_fix

* Update API.spec for reverse op
test=develop, test=document_fix

* fix Print, is_empty, read_file test=document_fix

* Refine the main comment of DynamicRNN.

* Refine the documentation of DynamicRNN's step_input function.

* Refine the documentation of DynamicRNN's static_input function.

* Refine the documentation of DynamicRNN's block function.

* Refine the documentation of DynamicRNN's memory function.

* Refine the documentation of DynamicRNN's update_memory and output function.

* Refine the code format and remove the method list.

* Refine the documentation of DynamicRNN's __call__ function.
test=develop
test=document_fix

* Minor modification.
test=develop
test=document_fix

* Fix some typo.

* Update API.spec.
test=develop
test=document_fix

* Refine the English according to the comments.

* Update API.spec.
test=develop
test=document_fix

* Fix some typo.

* Update API.spec.

* modify english api test=develop test=document_fix

- leaky_relu
- less_than
- log
- logical_and
- logical_or
- logical_xor
- logical_not

* fix some en doc test=develop test=document_fix

* fix new fluid.data API.spec test=develop test=document_fix

* fix typo test=develop, test=document_fix

* fix typo test=develop test=document_fix

* fix typo test=develop test=document_fix

* fix test=develop test=document_fix

* fix according review test=develop test=document_fix

* fix review test=develop test=document_fix

* fix according review advice test=develop test=document_fix

* fix the doc  test=develop test=document_fix

* fix API.spec test=develop test=document_fix

* fix English Doc of API:layers.py_func/sum, test=document_fix

* fix English Doc of API:layers.array_read/array_write/array_length,test=develop test=document_fix

fix reorder_lod_tensor_by_rank doc en 

* Refine StaticRNN api-doc-en

* refine static rnn doc

* refine api doc of fluid.data

* refine api doc from commends

* refine api.spec

* Refine grammer

* test=document_fix
Fix english doc api, invloves the op of greater_equal,greater_than,less_equal,not_equal,
rank,rsqrt,diag,linspace,reduce_all,reduce_any,sign,where,zeros_like,unique_with_counts.

* Fix some format problem in the op of sign and greather_than.
test=develop
test=document_fix

* Fix the example of zeros_like, and update api.spec
test=develop
test=document_fix

* optimize the error information when the input for while op has a wrong shape test=develop

* Fix recurrent op fails

For the variable used in outter block,
copy sub block's grad variable to outter block

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Fix unicode error

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Refine test code

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Fix seq2seq case fails

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* remove unreasonable code.

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Refine code according to comment

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* change the default value of summarize from -1 to 20 in Print op to improve ease of use, test=develop

* change the doc of API Print to make the document easier to understand, test=develop

* sequnce mask support max length tensor input; test=develop

* add rnn_impl.py; test=develop

* add basic gru lstm unittest; test=develop

* fix api spec; test=develop

* fix sequence_mask op bug;
test=develop
test=document_preview

* change +-*x to elmentwise_op; test=develop

* add mkl flag; test=develop

* fix rnn impl bug; test=develop

* update api spec; test=develop

* fix doc bug; test=develop

* fix lstm bugs; test=develop

refine StaticRNN api doc
test=develop
test=document_preview

Fix doc example for greater_equal, greater_than, less_equal, not_equal, rank, reduce_all, reduce_any, sign, where, diag (#18167)

* test=develop
fix greater_than, greater_equal, less_equal, not_equal, rank, reduce_all, reduce_any, sign, where, diag doc example

* test=develop
fix API.spec conflict

* fix code bug
test=develop

* fix API examples of assign, reverse and array_write
test=develop

* update API.spec
test=develop

* update API examples for array_length, array_read, array_write, assign, hard_sigmoid, hsigmoid, increment, ones, pow, reverse, uniform_random and zeros

* update API.spec for assign, reverse and array_write, etc.(#17287)
* test=develop

* API.spec test=develop

* update

* update test=develop

* update test=develop

* update

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* test=develop

* update

* update test=develop

* update test=develop

* fix:
1. infernce multi card occupy
2. facebox model inference occupy too much
test=develop

* refine python api comments: shuffle, while, scale, sampled_softmax_with_cross_entropy, scatter,
round, sin, sqrt, shape, split, soft_relu, slice, selu, ifelse, switch.
test=develodp

* fix conflict error.
test=develop

* update python APIs

test=document_preview test=develop

* update API.spec

test=document_preview

test=develop

* update merge_selected_rows

* update API.spec

test=document_preview

test=develop

* update API.spec

test=document_preview

test=develop

* fix the comment of less_than

test=develop

test=document_preview

* update API.spec

test=develop

test=document_preview

* update API.spec

test=develop

test=document_preview

* update API.spec

test=develop

test=document_preview

* update API.spec

test=develop

* update API

test=develop

* fix unique_name growth bug in dygraph mode,test=develop

* change generate_tmp to generate_with_ignorable_key,test=develop

* fix api doc according to the reviewer's comment test=develop

* Fix examples of fluid.layers.sums.
test=document_preview

* Correct the example of DynamicRNN and its functions.
test=develop

* Add 'import paddle.fluid as fluid' to examples.
test=develop

* Update API.spec.
test=develop

* Add space lines.
test=develop

* Update the API.spec.
test=develop

* Add api doc code examples

add or fix topk, squeeze, stack, StaticRNN,
StaticRNN memory in doc

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Add squeeze md5.

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Add import package

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* test=develop
update API.spec with hash, resolve conflict

* remove unused parameter