Adding a link to an adaptation of Google's word analogy task (15646 analogy questions) and pre-trained word embeddings:
--200-dimensional GloVe\
--300-dimensional CBOW and SkipGram\
--200-dimensional fastText (.text, .bin).
The training data (90.5mln tokens) included:
--Wikipedia\
--fiction texts taken from the open part of the EANC corpus\
--HC Corpora containing blogs and news articles collected by Hans Christensen in 2011\
--digitized and reviewed part of Armenian soviet encyclopedia from Wikisource\
--news articles.
Adding a link to an adaptation of Google's word analogy task (15646 analogy questions) and pre-trained word embeddings:
--200-dimensional GloVe\ --300-dimensional CBOW and SkipGram\ --200-dimensional fastText (.text, .bin).
The training data (90.5mln tokens) included:
--Wikipedia\ --fiction texts taken from the open part of the EANC corpus\ --HC Corpora containing blogs and news articles collected by Hans Christensen in 2011\ --digitized and reviewed part of Armenian soviet encyclopedia from Wikisource\ --news articles.