@inproceedings{zou-etal-2024-cqil,
    title = "{CQIL}: Inference Latency Optimization with Concurrent Computation of Quasi-Independent Layers",
    author = "Zou, Longwei  and
      Wang, Qingyang  and
      Zhao, Han  and
      Jiangangkong, Jiangangkong  and
      Yang, Yi  and
      Deng, Yangdong",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2024.acl-long.394/",
    doi = "10.18653/v1/2024.acl-long.394",
    pages = "7293--7307"
}