metadata
license: mit
Kalaallisut FastText Model
This is a test for FastText model creation that i use in different mini project with limited data.
This is an unsupervised FastText model trained on a corpus of Kalaallisut (Greenlandic) text data. The model was trained using the skip-gram method to learn high-quality word and sentence embeddings, capturing semantic relationships between words in the language.
Model Details
- Algorithm: FastText (Skip-gram)
- Vector Dimension: 300
- Training Data: 28,745 sentences (Kalaallisut text)
- Epochs: 10
- Minimum word occurrences: 2
Usage
You can load and use the model with FastText to obtain word embeddings, find similar words.