Hacker News new | ask | show | jobs
by doody_parizada 3108 days ago
We don't have words like v1agra. As I mentioned in the README, we took vectors pretrained on wikipedia. One of the possible improvements can be to train the vectors on our own dataset.