Datasets:
Missing features in the view of the dataset
#2
by
apohllo
- opened
Although the dataset contains the following features: generative_answer
and is_impossible
in the original files, the dataset python reader removes these features, when the dataset is loaded. Still these features are very important e.g. for generative question answering as well as detection of questions that don't have answers. Can you change the load script to include these features? It seems like a very simple thing to do... There's even a code that does that, but these lines are commented out. What is the rationale for not providing these features?