3 papers across 3 sessions
This paper fine-tunes inverse-folding model with feedbacks from folding model through DPO.