News for hackers. Stuff that matters.

A comment on Byte Latent Transformer: Patches Scale Better Than Tokens. Parent comment

Re:Byte Latent Transformer: Patches Scale Better Than Tokens by dv_dt on Saturday December 14, 2024 @11:57AM (#42418140)
So a variant might be to try using a some standard compression algorithm to train with?
[ Parent | on HN ]

There are no comments on this comment yet.