We provide multiple different models for NLP tasks on this website. Whether it is morphology, syntax or semantics that piques your interest, we’ve got something for you!
Rule-Based Morphology and Syntax
We compile every night the latest FSTs, CGs and dictionaries from the GiellaLT infrastructure. These are at the core of the uralicNLP Python library.
Finnish Semantic Relatedness Model
This model is a semantic model that captures the relatedness of Finnish words as word vectors. This model can be used in various tasks such as metaphor interpretation.
Finnish Sentiment Analysis Model
A model for a coarse sentiment analysis for Finnish. This model has been used poem generation.
Dictionaries for Uralic Languages
We have an online platform for browsing the GiellaLT dictionaries for multiple Uralic languages. We have been actively involved in authoring the Skolt Sami, Erzya, Moksha and Komi-Zyrian dictionaries. These dictionaries are also available in the uralicNLP Python library.
Nganasan and Kamas Speech Recognition
We have trained the state-of-the-art speech recognition models for Kamas and Nganasan.