F-coref: Fast, Accurate and Easy to Use Coreference Resolution
This work provides a faster coreference resolution tool for NLP practitioners, but it is incremental as it builds on existing architectures with optimizations.
The authors tackled the problem of slow coreference resolution by introducing F-coref, a fast model that processes 2.8K OntoNotes documents in 25 seconds on a V100 GPU with only a modest drop in accuracy compared to state-of-the-art models.
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. Our code is available at https://github.com/shon-otmazgin/fastcoref