S2O: Early Stopping for Sparse Attention via Online Permutation.
Yu Zhang, Songwei Liu, Chenqian Yan, Sheng Lin, Beichen Ning, Fangmin Chen, Xing Wang
Browse the full ACL paper archive.
Yu Zhang, Songwei Liu, Chenqian Yan, Sheng Lin, Beichen Ning, Fangmin Chen, Xing Wang
Browse the full ACL paper archive.