2. Fix loading checkpoint in filter pruning strategy by adding init_model option into config file.

* Fix name conficts in distillation.

* Download pretrain model in run_quant.sh

* Add demo for combined strategies.

* Fix quan.

* Restore args in run.sh

* Shift up fluid models

* Update nlp submodules

* Add redirect guide in org readme

* Move HiNAS_models to AutoDL

1. Remove pretrain directory in data
2. Add download scripts into run.sh

* Add demos for PaddleSlim on classification task.

* Refine structure of directory.