This model is part of the Text2Frappe pipeline designed for intelligent ERPNext query generation.
🔍 Task
Text2Text Generation (Field Selector)
Given a doctype, a natural language question, and a list of top fields (retrieved via SBERT ranking), this model selects the correct ERPNext field(s) that answer the question.
🧩 Example Input
Instruction: Select only the correct field(s) from the given top fields that answer the question.
Doctype: Sales Invoice Advance
Question: Which document type is most frequently associated with sales advance payments?
Top Fields: [reference_type, advance_amount, reference_name, exchange_gain_loss]
