1.2m English-language GloVe word embeddings trained on Twitter (50 dimensions)

A matrix of 1,193,514 word vectors (rows) and 50 dimensions (columns). GloVe embeddings trained on 2 billion tweets (27 billion tokens).

Format

A matrix of 1,193,514 rows and 50 columns

Source

https://nlp.stanford.edu/projects/glove/

References

Pennington, Jeffrey, Richard Socher, and Christopher Manning. (2014). "Glove: Global Vectors for Word Representation." Pp. 1532–43 in Proceedings of the 2014 conference on empirical methods in natural language processing

Examples

if (FALSE) { # \dontrun{

## download the model (once per machine)
download_pretrained("vecs_glove50_twitter")

## load the model each session
wv <- load_pretrained("vecs_glove50_twitter")

## check dims
dim(wv) == c(1193514, 50)

} # }