* fix sequence_pad cn doc

* add input output shape

* fix length info

* fix length info and pad_value info

* test=document_fix

* revision sequence_expand cn doc

* only surpport LoDTensor

* change name's text

* add input output shape

* add code example

* add code example

* test=document_fix

* fix sequence_expand_as cn doc

* only surpport LoDTensor

* add input output shape

* add code example

* add code example

* test=document_fix

change the output of retinanet_target_assign in api_cn

* fix apis, test=develop

* drop WeightedAverage_cn.rst, test=develop

* update layer api, test=develop

* fix dtype, test=develop

* polish en doc of data_loader

* follow comments

* use fluid.data()

* update cosine decay, test=develop

* update, test=document_fix

* update, test=document_fix

* update, test=document_fix

* update, test=document_fix

* update, test=document_fix

* update, test=document_fix

* update cn doc for create_py_reader_by_data

* refine decription and sample code for create_py_reader_by_data

* replace fluid.layers.data with fluid.data in the sample code of cn doc of create_py_reader_by_data op

* refine the description of capacity of create_py_reader_by_data

* fix description for create_global_var,create_parameter test=develop test=document_fix

update example code, use fluid.data swish, gaussian_random, thresholded_relu, test=document_preview (#1487)

* fix example for thresholded relu, test=document_preview

* update example code for swish, thresholded_relu, gaussian_random in chinese doc, test=document_preview

* fix example code for thresholded_relu, test=document_preview

* fix ref for name, test=document_preview

updates sequence_scatter_cn.rst 

updates sequence_unpad_cn.rst

* refine build and install tutorial doc for paddle inference

* remove trt_benchamrk after merge it to trt doc

update sequence_softmax_cn.rst

* add astype api in Variable and refine code style

* add astype api in Variable and refine code style

* add astype api in Variable and refine code style

* refine doc

* polish cn document of unique_name

* polish guard and switch

* add example code

* add filter to simplify optimizer api doc

* update related optimizer doc

* refine prepare steps doc, test=develop

* replace layers.data with fluid.data, test=develop

* update softmax_cn.rst, test=develop

* update softmax_cn.rst, test=develop

* fix error, test=develop

* add comment for name, test=develop

* update softmax_cn.rst, test=develop

* update softmax_cn.rst, test=develop

* fix softmax_cn.rst, test=develop

* fix softmax_cn.rst, test=develop

* polish model average zh doc api

* polish details

* polish based review commment

* fix indent

* fix the doc format of io.chain.

test=develop
test=document_preview

* fix fluid.data code example, test=develop, test=document_preview

* skip PipeReader in IO

* skip PipeReader in IO
test=develop

* fixed bugs about PipeReader

change fluid.layers.data to fluid.data in api_cn for retinanet_target_assign, sigmoid_focal_loss, retinanet_detection_output. (#1473)

* test=document_preview
change  to  in api_cn for retinanet_target_assign, sigmoid_focal_loss, retinanet_detection_output.

* test=document_preview
remove Notice: this OP supports CPU mode only in api_cn for retinanet_target_assign and retinanet_detection_output