LP model without templet

[Peter-zds]

Could you please provide more details about how to implement the works mentioned in the paper :
"Fine-tuning a pre-trained BART (Lewis et al., 2020) model, we directly learned a mapping from a high-level instruction to a sequence of subtasks (e.g. “Drop a clean pan on the table” → “(PickupObject, Pan), (PutObject,Sink), ...”). "
I would be very grateful if you could provide the code or the details of implementation.