@inproceedings{tricksfortrainingsparsetranslationmodels,
  title = {Tricks for Training Sparse Translation Models},
  author = {Dheeru Dua and Shruti Bhosale and Vedanuj Goswami and James Cross and Mike Lewis and Angela Fan},
  year = {2021},
  booktitle = {NAACL 2022 7},
  eprint = {2110.08246},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2110.08246v1},
}