NCRF++: An Open-source Neural Sequence Labeling Toolkit
This toolkit addresses the need for accessible and efficient neural sequence labeling tools for NLP researchers and practitioners, though it is incremental as it builds on existing methods.
The authors developed NCRF++, an open-source toolkit for neural sequence labeling that enables quick implementation of models with CRF inference layers through configuration files, built on PyTorch for GPU-accelerated efficiency and including implementations of state-of-the-art models like LSTM-CRF.
This paper describes NCRF++, a toolkit for neural sequence labeling. NCRF++ is designed for quick implementation of different neural sequence labeling models with a CRF inference layer. It provides users with an inference for building the custom model structure through configuration file with flexible neural feature design and utilization. Built on PyTorch, the core operations are calculated in batch, making the toolkit efficient with the acceleration of GPU. It also includes the implementations of most state-of-the-art neural sequence labeling models such as LSTM-CRF, facilitating reproducing and refinement on those methods.