marcelroed on GigaToken: ~1000x faster Language model tokenization · Hackerly

Skip to content

Top New Best Ask Show Jobs

Top New Best Ask Show Jobs

Comment on GigaToken: ~1000x faster Language model tokenization parent

Comments

Running the numbers now

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.