te-sla/FastTextSr

FastText Sr
Обучаван над корпусом српског језика - 9.5 милијарди речи Међу датотекама се налазе модели у Gensim, али и оригиналном формату	Trained on the Serbian language corpus - 9.5 billion words The files include models in both Gensim and the original format.

from gensim.models import FastText
model = Word2Vec.load("TeslaFT")
examples = [
    ("dim", "zavesa"),
    ("staklo", "zavesa"),
    ("ormar", "zavesa"),
    ("prozor", "zavesa"),
    ("draperija", "zavesa")
]
for e in examples:
    model.wv.cosine_similarities(ft.wv[e[0]], ft.wv[[e[1]]])[0]

0.5305264
0.7095266
0.6041575
0.5771946
0.8870213

from gensim.models.fasttext import load_facebook_model
model = load_facebook_model("TeslaFT.bin")
examples = [
    ("dim", "zavesa"),
    ("staklo", "zavesa"),
    ("ormar", "zavesa"),
    ("prozor", "zavesa"),
    ("draperija", "zavesa")
]
for e in examples:
    model.wv.cosine_similarities(ft.wv[e[0]], ft.wv[[e[1]]])[0]

0.5305264
0.7095266
0.6041575
0.5771946
0.8870213

Author

Mihailo Škorić

@procesaur

Computation

TESLA project

@te-sla

@inproceedings{stankovic-dict2vec,
  author    = {Ranka Stanković, Jovana Rađenović, Mihailo Škorić, Marko Putniković},
  title     = {Learning Word Embeddings using Lexical Resources and Corpora},
  booktitle   = {15th International Conference on Information Society and Technology, ISIST 2025, Kopaonik},
  year      = {2025},
  address = {Kopaonik, Belgrade}
  publisher = {SASA, Belgrade},
  url       = {https://doi.org/10.5281/zenodo.15093900}
}

Истраживање jе спроведено уз подршку Фонда за науку Републике Србиjе, #7276, Text Embeddings – Serbian Language Applications – TESLA

This research was supported by the Science Fund of the Republic of Serbia, #7276, Text Embeddings - Serbian Language Applications - TESLA