It's a great video, but I still haven't figured out how to apply it... Do you need to use their work to evaluate the model, see what types of questions it is wrong on, and, for example, tune it to improve it for them, or how? Can anyone clarify?
It's a great video, but I still haven't figured out how to apply it... Do you need to use their work to evaluate the model, see what types of questions it is wrong on, and, for example, tune it to improve it for them, or how? Can anyone clarify?