imdbo's picture
Update README.md
79bd511
|
raw
history blame
1.93 kB
metadata
license: mit

license: mit

Descrici贸n do Modelo

Modelo feito con OpenNMT para o par espa帽ol-galego utilizando unha arquitectura transformer.

Como utilizar

onmt_translate -src input_text -model NOS-MT-es-gl -output ./output_file.txt -replace_unk -phrase_table phrase_table-es-gl.txt -gpu 0
  • O resultado da traduci贸n estar谩 no PATH indicado no flag -output.

Adestramento

Datos utilizados para o adestramento

Aut茅nticos e Sint茅ticos (Translitera莽茫o)[Colocar Paper]

Procedemento de adestramento

Tokenization feita co tokenizador de linguakit https://github.com/citiususc/Linguakit

BPE

Hiperpar谩metros

Colocar o yaml para cada um dos pares

Avaliaci贸n A avalaci贸n dos modelos 茅 feita cunha mistura de tests desenvolvidos internamente (gold1, gold2, test-suite) con outros datasets dispon铆beis en galego (Flores).

GOLD 1 GOLD 2 FLORES TEST-SUITE
36.8 47.1 32.3 42.7

Informaci贸n adicional

Licensing information

Apache License, Version 2.0

Financiamento

Financiamento

This research was funded by the project "N贸s: Galician in the society and economy of artificial intelligence", agreement between Xunta de Galicia and University of Santiago de Compostela, and grant ED431G2019/04 by the Galician Ministry of Education, University and Professional Training, and the European Regional Development Fund (ERDF/FEDER program), and Groups of Reference: ED431C 2020/21.

Citation Information

@article{, title={}, author={}, year={2022}, url={} }