@inproceedings{lightseqahighperformanceinferencelibrary, title = {LightSeq: A High Performance Inference Library for Transformers}, author = {Xiaohui Wang and Ying Xiong and Yang Wei and Mingxuan Wang and Lei LI}, year = {2020}, booktitle = {NAACL 2021 4}, eprint = {2010.13887}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2010.13887v4}, }