News for hackers. Stuff that matters.
Re:A Multimodal Dataset with One Trillion Tokens by 0xDEADFED5 on Wednesday July 24, 2024 @08:01PM (#41063501)
Salesforce produced one of the best Llama-3(8B) finetunes, IMO: SFR-Iterative-DPO-LLaMA-3-8B-R

Hopefully they do something with Llama-3.1

There are no comments on this comment yet.