This paper describes the design and use of the graph-based parsing framework and toolkit UniParse, released as an open-source python software package developed at the IT University, Copenhagen Denmark. UniParse as a framework novelly streamlines research prototyping, development and evaluation of graph-based dependency parsing architectures. The system does this by enabling highly efficient, sufficiently independent, readable, and easily extensible implementations for all dependency parser components. We distribute the toolkit with ready-made pre-configured re-implementations of recent state-of-the-art first-order graph-based parsers, including highly efficient Cython implementations of feature encoders and decoding algorithms, as well as off-the-shelf functions for computing loss from graph scores.
展开▼
