NCRF++: An Open-source Neural Sequence Labeling Toolkit

TL;DR

NCRF++ is an open-source toolkit built on PyTorch that enables quick development and experimentation with neural sequence labeling models, featuring flexible configurations and GPU acceleration.

Contribution

It introduces a flexible, easy-to-use toolkit for neural sequence labeling that supports various models and simplifies model customization and reproduction.

Findings

Supports multiple neural sequence labeling models

Efficient batch processing with GPU acceleration

Facilitates model reproduction and refinement

Abstract

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.