A copy of the [NextQA dataset](https://arxiv.org/abs/2105.08276), used to demonstrate [fine-tuning Aria on video dataset](https://github.com/rhymes-ai/Aria/tree/main/examples).