gama-verse data

Large files used by gama-verse (GAMA, Galician Metrical Analyzer) for light-weight normalization of 19th-century Galician poetry. This does not perform full normalization to a current Galician norm, but rather light-weight edits that preserve words' metrically relevant features (e.g. number of syllables), approximate orthographic stress to the current norm, and disambiguate some stress cases to help scansion. The files are downloaded with the gama-verse-download command; see the gama-verse README.

File Size Content License
nos-127.klm.bin 736 MB KenLM language model for Galician, trained on the Nós corpus CC BY 4.0
gl-vocab.pkl 121 MB Galician vocabulary, based on Linguakit and Apertium GPLv3
aspell-es-expanded.txt.pkl 13 MB Spanish vocabulary: the word forms of the aspell Spanish dictionary (COES, Santiago Rodríguez and Jesús Carretero), expanded GPL-2.0-or-later

The .pkl files are Python pickles (sets of word forms).

Citation

If you use these files, please cite the paper where the normalization is described:

Pablo Ruiz Fabo, Pauline Moreau, Anxo Alonso Pérez. (2026). Automatic Metrical Scansion of Galician Poetry: First Results. Proceedings of PROPOR - 17th International Conference on Computational Processing of Portuguese.

The language model was trained on CorpusNÓS:

Iria de-Dios-Flores, Silvia Paniagua Suárez, Cristina Carbajal Pérez, Daniel Bardanca Outeiriño, Marcos Garcia, Pablo Gamallo. (2024). CorpusNÓS: A massive Galician corpus for training large language models. Proceedings of the 16th International Conference on Computational Processing of Portuguese - ACL Anthology (Volume 1), 593-599.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support