Abstract
We introduce fastcoref, a python package for fast, accurate, and easy-to-use English coreference resolution. The package is pip-installable, and allows two modes: an accurate mode based on the LINGMESS architecture, providing state-of-the-art coreference accuracy, and a substantially faster model, F-COREF, which is the focus of this work. F-COREF allows to process 2.8K OntoNotes documents in 25 seconds on a V100 GPU (compared to 6 minutes for the LINGMESS model, and to 12 minutes of the popular AllenNLP coreference model) with only a modest drop in accuracy. The fast speed is achieved through a combination of distillation of a compact model from the LingMess model, and an efficient batching implementation using a technique we call leftover batching.
| Original language | English |
|---|---|
| Title of host publication | System Demonstrations |
| Editors | Wray Buntine, Maria Liakata |
| Publisher | Association for Computational Linguistics (ACL) |
| Pages | 48-56 |
| Number of pages | 9 |
| ISBN (Electronic) | 9781955917551 |
| DOIs | |
| State | Published - 2022 |
| Event | 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, AACL-IJCNLP 2022 - Virtual, Online Duration: 20 Nov 2022 → 23 Nov 2022 |
Publication series
| Name | Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing: Long Paper, AACL-IJCNLP 2022 |
|---|---|
| Volume | 4 |
Conference
| Conference | 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, AACL-IJCNLP 2022 |
|---|---|
| City | Virtual, Online |
| Period | 20/11/22 → 23/11/22 |
Bibliographical note
Publisher Copyright:©2022 Association for Computational Linguistics.