Rust interface to word2vec word vectors.
This crate provides a way to read a trained word vector file from word2vec. It doesn't provide model training and hence requires a already trained model.
Documentation is available at https://github.com/DimaKudosh/word2vec/wiki
Add this to your cargo.toml
:
[dependencies]
# …
word2vec = "0.3.3"
Example for word similarity and word clusters:
extern crate word2vec;
fn main(){
let model = word2vec::wordvectors::WordVector::load_from_binary(
"vectors.bin").expect("Unable to load word vector model");
println!("{:?}", model.cosine("snow", 10));
let positive = vec!["woman", "king"];
let negative = vec!["man"];
println!("{:?}", model.analogy(positive, negative, 10));
let clusters = word2vec::wordclusters::WordClusters::load_from_file(
"classes.txt").expect("Unable to load word clusters");
println!("{:?}", clusters.get_cluster("belarus"));
println!("{:?}", clusters.get_words_on_cluster(6));
}