Hi,
I’m currently working on a NLP project that involves biomedical event extraction, I am using the Genia dataset (2011). My team decided to work on various approaches, my job is to try zero-shot and few-shot on Flan-T5. However, due to the limit of LM, I suppose, the model did not work efficiently. Did anyone have tried this way or do you have any suggestion of other approaches?
elaborate limitations of your LM you are stating, and share screenshots of your results