GigaToken: ~1000x faster Language model tokenization1by Hacker News1w agogithub.com fmode.me/n/gigatoken-1000x-faster-language-model-tokenization Read the full article https://github.com/marcelroed/gigatoken/More from FM NewsRunning Kimi K3 on MI355X at Better Performance per Dollar Than B3001 pts · Hacker NewsElena, a library for building Progressive Web Components1 pts · Hacker NewsThe teenage hitmen recruited to kill across Europe1 pts · Hacker NewsZinley1 pts · Product Hunt Back to FM News