added tests and test workflow, small cleanup

implemented option to use inference endpoints, implemented parameter selection, updated UI, cleaned up return formats of models

removed force pushed, load models separately to optimize ram usage

renamed poc_app.py to app.py, clean up for app logic, implemented function to yield intermediate results, added generation length parameter to config, changed path handling in app.py

added attribution example, created local pipelines, implemented poc_app.py as a proof of concept