* refine dist train RPC deps

* clean up

* clean up

* fix ut

* remove input for fetch_barrier

* follow comments

Add concat op converter

* test fix release branch api check

* fix reviews > 30

* check approval after test, check api diff before test

Add unstack op

Enforce requested size of tensor to be sufficiently large

Reusing primitives for forward Batch Norm operator

Fix random diff between python2 and python3

Feature fusion expand concat fc

Fix concat bug

* Support int8 type.

* Add pad_constant_batch_size_like

* refine pad_op

* optimize memory

* platform module

* Update profiler.h

refine fusion lstm

* memory module

* "fix ci"

* fix auc layer and add check for auc op

* use input to check if states are inited

* optimize code

Speedup stack_op