* fix ptb reader and remove useless language model; test=develop

* add ptb reader; test=develop

* Add usage for distribute Light-NAS searching.

* Refine usage doc of PaddleSlim.

* using new load api

* test=develop, add license

* fix2744

* remove useless comment

* make code less ugly

* optimize error message when result is none

add copyright for cycle_gan and se_resnext

* fix ptb author bug; test=develop

* fix ptb author bug; test=develop

test=develop

fix ceroot path bug

* fix save_inference_model cannot find feed vars

* remove comment

* fix format

* add comment for pruned var

* add log for prune

* Update mnist_dygraph.py

fix bug

* add muti card support for se_resnext

* add some description to readme.md

* add ce for cyclegan

* fix code style

* add ce for ptb_lm

* fix resnet for resnet34

* fix ResNet18 and ResNet34

* update modelzoo and fix configs

* fix configs

* fix resnet bn

* update readme

* rm useless png

* fix bn warning

Fix the leaving out of subprocess importing.

* Clean up a bit

* Add `__shared__` annotation for shared variables between modules

* Dedup `num_classes` configurations

* Document `__shared__` usage

* Remove some unused variable

* Improve docstring and add comments for `__shared__`

* add cn doc for INSTALL/GETTING_START

* cn for GETTING_STARTED

* fix doc

add run shell and infer shell for cgan and dcgan
update infer.py for gan and dcgan

* add instance norm

* Add a model for WebVision 2018.

* baiduzhixing score tools

* update name and add baseline model

* Update mnist_dygraph.py

fix bug

* add ce to se_resnext

* delete useless comments and fix unique_name bugs

* add sentiment ce

* delete some comment

* fix code style

* Update mnist_dygraph.py

fix bug

* add muti card support for se_resnext

* add some description to readme.md

some of the global config values are reference type, e.g., objects, lists or
dicts, if the created module instances modify those values (see FPN and
spatial_scale), global config will reflect these changes, and instances of the
same class created later will inherit the changed values

* refine bbox_normalize in infer.py

* add is_bbox_normalize

* rename _forward to build

* check callable