Back to News
Advertisement
Advertisement

Discussion (1 Comments)Read Original on HackerNews

ricardobeatabout 4 hours ago
Love this kind of experiment. Would the model perform better with word tokens?