Skip to main content


If the rumors are true, this week could be another exciting week for opensource LLMS! Meta may release Llama-3-405b on this Tuesday. Also there could be updates to 8b and 70b models distilled from 405B. Joe Spisak, a product director at Meta says they were initially going to call Llama 3 8b and 70b a prerelease or preview because these models didn't have all the things they planned to release.
Sources:
theinformation.com/briefings/m…
x.com/AlpinDale/status/1814814…
youtu.be/r3DC_gjFCSA?feature=s…
#LLM #AI #ML
#AI #ML #llm
in reply to Chi Kim

I'm hoping for Meta to release something multi-modal? Like an upgarade to Llama3 or something.
in reply to victor tsaran

@vick21 Meta released a small multimodal called Chameleon a month ago, but who knows maybe llama-3-405b is multimodal? about.fb.com/news/2024/06/rele…
in reply to Chi Kim

Chameleon does not run in Ollama env, correct? BTW, I do hope for a multi-modal in Llama though. They really need to catch up with Open AI! :)
in reply to victor tsaran

@vick21 No, it doesn't work. I lost faith in multimodal on llama.cpp, which Ollama relies on. Llama.cpp dropped multimodal support on March 7 (supposedly temporarily), but there is no momentum or interest among the devs to bring it back. They all seem preoccupied with providing immediate next day support for new text-only LLMs being released constantly.
in reply to Chi Kim

Interesting. I'm sort of getting tired of all these models being released every day. It increasingly looks more and more like a gaming exercise! :)