sagorsarker's picture
Update README.md
9a544eb
metadata
license: mit

Bangla Glove Vectors

This is a collection of pre-trained glove vectors for the Bengali language. You can find details training in this repository.

This model is build for bnlp package.

Datasets

Model Details

  • wikipedia+crawl_news_articles (39M(39055685) tokens, 0.18M(178152) vocab size, 195.4MB download): bn_glove.39M.300d.zip

  • wikipeida+crawl_news_articles (39M(39055685) tokens, 0.18M(178152) vocab size, 100d vectors, 65MB download): bn_glove.39M.100d.zip

  • Wikipedia (20M(19965328) tokens, 0.13M(134255) vocab, 300d vectors, 145.9MB download): bn_glove.300d.zip

  • Wikipedia (20M(19965328) tokens, 0.13M(134255), 100d vectors, 51.2MB download): bn_glove.100d.zip

Usage

from bnlp import BengaliGlove
glove_path = "bn_glove.39M.100d.txt"
word = "গ্রাম"
bng = BengaliGlove()
res = bng.closest_word(glove_path, word)
print(res)
vec = bng.word2vec(glove_path, word)
print(vec)