GigaToken: ~1000x faster Language model tokenization

(github.com)

167 points | by syrusakbary  3 hours ago

28 comments