emLam -- a Hungarian Language Modeling baseline
arXiv:1701.07880
Abstract
This paper aims to make up for the lack of documented baselines for Hungarian language modeling. Various approaches are evaluated on three publicly available Hungarian corpora. Perplexity values comparable to models of similar-sized English corpora are reported. A new, freely downloadable Hungar- ian benchmark corpus is introduced.
Additional resources: - the emLam repository: https://github.com/DavidNemeskey/emLam - the emLam corpus: http://hlt.bme.hu/en/resources/emLam