News for hackers. Stuff that matters.
Re:33TB of text data for a 1T-parameter model by qumpis on Sunday April 16, 2023 @06:02AM (#35588660)
Train only on data that is hard to predict/complete?

There are no comments on this comment yet.