Currently, each approach for text generation results in different structures for the question data generated. The text file generated from each generation model should have the same schema to make it easier to compare and validate the questions per approach. This will make evaluation and testing more streamlined as well.
Currently, each approach for text generation results in different structures for the question data generated. The text file generated from each generation model should have the same schema to make it easier to compare and validate the questions per approach. This will make evaluation and testing more streamlined as well.