▲ 196 ▼ Smaug-72B-v0.1: The New Open-Source LLM Roaring to the Top of the Leaderboard (huggingface.co) submitted 2 years ago by hexual@lemmy.world to c/technology@lemmy.world 60 comments fedilink hide all child comments Abacus.ai: We recently released Smaug-72B-v0.1 which has taken first place on the Open LLM Leaderboard by HuggingFace. It is the first open-source model to have an average score more than 80.
[–] FenrirIII@lemmy.world 2 points 2 years ago (1 child) OOTL: What is a LLM and what does it do? permalink fedilink source hideshow 2 child comments replies: [–] saltesc@lemmy.world 22 points 2 years ago (1 child) Large Language Model AI. Like ChatGPT. permalink fedilink source parent hideshow 2 child comments replies: [–] FaceDeer@kbin.social 0 points 2 years ago (1 child) And at 72 billion parameters it's something you can run on a beefy but not special-purpose graphics card. permalink fedilink source parent hideshow 2 child comments replies: [–] glimse@lemmy.world 6 points 2 years ago (2 children) Based on the other comments, it seems like this needs 4x as much ram than any consumer card has permalink fedilink source parent hideshow 4 child comments replies: [–] FaceDeer@kbin.social 4 points 2 years ago It hasn't been quantized, then. I've run 70B models on my consumer graphics card at a reasonably good tokens-per-second rate. permalink fedilink source parent [–] DarkThoughts@fedia.io 2 points 2 years ago I'm curious how local generation goes with potentially dedicated AI extensions using stuff like tensor cores and their own memory instead of hijacking parts of consumer GPUs for this. permalink fedilink source parent
[–] saltesc@lemmy.world 22 points 2 years ago (1 child) Large Language Model AI. Like ChatGPT. permalink fedilink source parent hideshow 2 child comments replies: [–] FaceDeer@kbin.social 0 points 2 years ago (1 child) And at 72 billion parameters it's something you can run on a beefy but not special-purpose graphics card. permalink fedilink source parent hideshow 2 child comments replies: [–] glimse@lemmy.world 6 points 2 years ago (2 children) Based on the other comments, it seems like this needs 4x as much ram than any consumer card has permalink fedilink source parent hideshow 4 child comments replies: [–] FaceDeer@kbin.social 4 points 2 years ago It hasn't been quantized, then. I've run 70B models on my consumer graphics card at a reasonably good tokens-per-second rate. permalink fedilink source parent [–] DarkThoughts@fedia.io 2 points 2 years ago I'm curious how local generation goes with potentially dedicated AI extensions using stuff like tensor cores and their own memory instead of hijacking parts of consumer GPUs for this. permalink fedilink source parent
[–] FaceDeer@kbin.social 0 points 2 years ago (1 child) And at 72 billion parameters it's something you can run on a beefy but not special-purpose graphics card. permalink fedilink source parent hideshow 2 child comments replies: [–] glimse@lemmy.world 6 points 2 years ago (2 children) Based on the other comments, it seems like this needs 4x as much ram than any consumer card has permalink fedilink source parent hideshow 4 child comments replies: [–] FaceDeer@kbin.social 4 points 2 years ago It hasn't been quantized, then. I've run 70B models on my consumer graphics card at a reasonably good tokens-per-second rate. permalink fedilink source parent [–] DarkThoughts@fedia.io 2 points 2 years ago I'm curious how local generation goes with potentially dedicated AI extensions using stuff like tensor cores and their own memory instead of hijacking parts of consumer GPUs for this. permalink fedilink source parent
[–] glimse@lemmy.world 6 points 2 years ago (2 children) Based on the other comments, it seems like this needs 4x as much ram than any consumer card has permalink fedilink source parent hideshow 4 child comments replies: [–] FaceDeer@kbin.social 4 points 2 years ago It hasn't been quantized, then. I've run 70B models on my consumer graphics card at a reasonably good tokens-per-second rate. permalink fedilink source parent [–] DarkThoughts@fedia.io 2 points 2 years ago I'm curious how local generation goes with potentially dedicated AI extensions using stuff like tensor cores and their own memory instead of hijacking parts of consumer GPUs for this. permalink fedilink source parent
[–] FaceDeer@kbin.social 4 points 2 years ago It hasn't been quantized, then. I've run 70B models on my consumer graphics card at a reasonably good tokens-per-second rate. permalink fedilink source parent
[–] DarkThoughts@fedia.io 2 points 2 years ago I'm curious how local generation goes with potentially dedicated AI extensions using stuff like tensor cores and their own memory instead of hijacking parts of consumer GPUs for this. permalink fedilink source parent