* Revert back to `fluid.clip` for gradient clipping

current API is in flux, wait for it to stabilize

* Please CI

still buggy

* support face detection in python inference

* support face detection in python deploy

* add devices_num on reader

* refine capacity value

* refine bufsize

* update cpp_infer for paddle latest

* polish code

* export config file in export_model.py

* add custom label_list & dump_box

* add infer_cfg.yml & doc

* export label_list to inference config

* update PaddleTensor to ZeroCopyTensor

* Update README.md
* Update change logs
* Update docs/advanced_tutorials/inference/INFERENCE.md

* add padstride in cpp_infer and refine doc
* refine doc for export_model

add libra rcnn, including
  * libra loss
  * libra fpn(bfp)
  * libra sampling

* add blazeface nas demo

* fix format

* fix configs

* fix reduce_rate

* refine code & doc for cpp_infer

* add cython for pycocotools install

* load inference

* add doc

* add config file for tensorrt_infer

* add padding for rcnn preprocess

* refine code & doc

* Add GIOU loss.
* Add DIOU/CIOU loss.
* Change loss box type in bbox_head.py.
* Remove fuse_elewise_add_act_ops flag in train.py (Error may happen for elem_divbased on the flag).

- 1.add cn docs
- 2.fix fddb evaluate
- 3.fix get_shrink
- 4.delete FLAGS.dataset_dir

* split yolov3 loss

* Polish Reader to simplify preprocessing logic.
#  sample_transforms-> make batch -> batch_transforms in Reader.
* Clean some code
* Imporve yolov3_r50vd_dcn_obj365_pretrained_coco 41.4 to 41.8.
* Update all configs.

* Add cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms model on oidv5 and obj365.
* Add obj365 config annotation
* Update docs.

* update code & doc

to avoid cycle import when required packages are not installed

* add export_model doc

* Error out if fp16 is enabled and affine channel is used

* Add more documentation on mixed precision training