As the title.

* fix add grad bug test=develop

* update style test=develop

* refine grad maker, test=develop

* refactor tracer stage 1, test=develop

* merge develop to solve conflict third times, test=develop

* fix API param bug of recompute.backward(), test=develop

* fix fc padding during fusion, test=develop

* fix optim model inference after SaveOptimModel, test=develop

* fix communicator when breaking under PyReader mode, test=develop
* revert some vlog level to 0, test=develop

add lookup_table_dequant_op

As model fails when enable int8 quant, so disable allocate memory in cpu
for small variable.

* change the ci trt from version 5. to 6.0

* paddle-trt dynamic shape support init

* conv+bias or conv+bn dynamic shape support
test=develop

* modity trt engine opconvert
test=develop

* fix ci error
test=develop

* refine the profiler print test=develop

* support basic tensor.shape. 

* Support tensor.shape with dependencies. 

* polish detail implement of data loader, test=develop

* solve coverage ci problem, test=develop

In elementwise_min api doc has the error, all the api just use the elementwise_max op, we just correct the doc.

1. copy.deepcopy in NameVisitor should be changed to copy.copy to make hash or set work
2. read_context should be type of gast.Load()/gast.AugLoad(), not gast.Load/gast.AugLoad

fix concat_mkldnn op when encounter extreame conditions.

* reduce default attrs for dynamic graph, test=develop

* add some explanations for explicit attr, test=develop

* tweak explicit attr comments, test=develop

Add basic support for while in translating dygraph to static

1. Analysis the variable liveness in class NameVisitor
2. Replace while key word using while_loop API

* Added an option to use external FP32 model in QAT comparison test

test=develop

* Modify the way of inserting newly defined func_nodes test=develop

* fix loaded program load bug; test=develop

* first version

* speed backward engin; test=develop

* remove useless code; test=develop

* reconvery io.py; test=develop

* remove useless code; test=develop

* remove useless code; test=develop

* add recorded cuda memory apis, fix typo, test=develop

* add more ut, test=develop

* follow comments, test=develop

* fix py35 incompatible issues, test=develop

* encapsulate the PaddleTensorToLoDTensor, test=develop

* serialize the pd_tensor, test=develop

* serialize tensors to file, test=develop

* add fluid.device_guard to specify the device type for Op