Je vous transmets un message de Mathieu Valette qui renseigne sur la plateforme MultiTAL qui recense des outils pour annoter, segmenter, etc.
The MultiTAL digital platform is now up and running at http://multital.inalco.fr .
MultiTAL is a curated collection of NLP tools that specifically focuses on either Oriental or less-resourced languages. It makes it possible to easily select and run a specific tool in accordance with such criteria as goal, language, OS or methodology…
Each tool is offered along with its own standardised multilingual documentation -different from the tool developer’s- which describes how to install and run the tool.
An ontology related to the different tools can also be SPARQL-queried or downloaded (RDF/XML format).
At time of writing 163 tools have been tested, 91 of which have been retained. They can run on 44 languages, even though our tests have been conducted on Arabic, Chinese, Hindi, Japanese and Russian.
Work on the MultiTAL platform is still in progress, please let us know how it can be improved.
The MultiTAL platform has been funded by USPC and has been designed by ER-TIM, a research team of INaLCO.
OpenEdition vous propose de citer ce billet de la manière suivante :
Eva Schaeffer-Lacroix (31 janvier 2017). Plateforme MultiTAL. Club Corpus. Consulté le 13 décembre 2024 à l’adresse https://doi.org/10.58079/mzao