# MoViE-MCAN (VQA 2020 Challenge Winner)

This repository contains the code for pytorch implementation of MoViE+MCAN model, winner of the VQA Challenge 2020 at CVPR. Please cite the following papers if you are using MoViE+MCAN model from mmf:

* Nguyen, D. K., Goswami, V., & Chen, X. (2020). *Revisiting Modulated Convolutions for Visual Counting and Beyond*. arXiv preprint arXiv:2004.11883. ([arXiV](https://arxiv.org/abs/2004.11883))
```
@article{nguyen2020revisiting,
  title={Revisiting Modulated Convolutions for Visual Counting and Beyond},
  author={Nguyen, Duy-Kien and Goswami, Vedanuj and Chen, Xinlei},
  journal={arXiv preprint arXiv:2004.11883},
  year={2020}
}
```

and

* Jiang, H., Misra, I., Rohrbach, M., Learned-Miller, E., & Chen, X. (2020). *In Defense of Grid Features for Visual Question Answering*. arXiv preprint arXiv:2001.03615. ([arXiV](https://arxiv.org/abs/2001.03615))
```
@article{jiang2020defense,
  title={In Defense of Grid Features for Visual Question Answering},
  author={Jiang, Huaizu and Misra, Ishan and Rohrbach, Marcus and Learned-Miller, Erik and Chen, Xinlei},
  journal={arXiv preprint arXiv:2001.03615},
  year={2020}
}
```

Please see [https://mmf.sh/docs/projects/movie_mcan](https://mmf.sh/docs/projects/movie_mcan) for more details on how to use MoViE+MCAN model.