* 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