Can we have a function that is similar to the fasttext `print-n-grams`. I personally need it to learn the embedding of bi-grams, related to a word