* bug fix: invalid learning rate decay in pserver async mode

* Fix transpiler en doc (#20149)

* test=develop,test=document_fix,fix transpiler doc,add API.spec

* test=develop,test=document_fix,fix transpiler doc,add API.spec

add executor.FetchHandler for train/infer from the dataset

* Speed GEO-SGD (#20158)

* delete debug vlog & add rpc function & fix word2vec bug & speed GEO-SGD

Heartbeat for distributed async training.

* refector geo sgd & communicator

* add a base class for the Communicator
* add AsyncCommunicator Impl for async distributed training

* rpc retry for asycsend/get/prefetch

* test=develop, change retry vlog level to 3

* test=develop, set default grpc_retry_times is 3

* remove -Wmaybe-uninitialized warning

test=develop

* remove uninitialized op_handle_ in scale_loss_grad_op_handle.cc

test=develop

* test=develop,  communicator merge add => merge average

* fix correctness of the communicator

* fix a bug in send thread when sending var context is empty, test=develop

* add lookup_table_prefetch_op and prefetch optimize, test=develop

* remove remote prefetch GPU supported

* word2vec force with CPU, test=develop

* test dist remote lookup table force with CPU, test=develop

Unset unittests http_proxy env to avoid timeout.

Polish fleet API to support cuda collective mode and nccl2 mode

* add is_runnning in communicator, test=develop

* optimize communicator merge sparse gradient test=develop

* revert multithread selected rows merge add test=develop

* follow comment test=develop

* fix save/load in Fleet
* add UT framework of Fleet

* remove attribute in Allocator::Allocate, test=develop

* fix travis ci error, test=develop

Async exe support communicator

test=develop

* optimize communicator flag

* change flags in init py test=develop

test=develop