Support more QA tasks?

[pfliu-nlp]

• [ ] NQ (https://ai.google.com/research/NaturalQuestions/dataset)
• [ ] TriviaQA (https://nlp.cs.washington.edu/triviaqa/) datasets
• [ ] HotpotQA
• [ ] DROP

[divija96]

Mentioning other popular QA datasets here for reference:

• [ ] WebQA: Multimodal hence might be difficult to evaluate.
• [ ] BioASQ: Biomedical semantic indexing.