* add correctness test on arc for llama model
* modify layer name
* add falcon ut
* refactor and add ut for falcon model
* modify lambda positions and update docs
* replace loading pre input with last decodelayer output
* switch lower bound to single model instead of using the common one
* make the code implementation simple
* fix gpu action allocation memory issue
* Add gpu workflow and a transformers API inference test
* Set device-specific env variables in script instead of workflow
* Fix status message
---------
Co-authored-by: sgwhat <ge.song@intel.com>
