ryujin470@fedia.io to Technology@beehaw.org · 2 months agoOpenAI releases a free GPT model that can run on your laptopwww.theverge.comexternal-linkmessage-square8fedilinkarrow-up138arrow-down17file-text
arrow-up131arrow-down1external-linkOpenAI releases a free GPT model that can run on your laptopwww.theverge.comryujin470@fedia.io to Technology@beehaw.org · 2 months agomessage-square8fedilinkfile-text
minus-squareNate@piefed.alphapuggle.devlinkfedilinkEnglisharrow-up14·2 months ago*if you have a laptop with 16gb of vram. Otherwise you’ll be watching ollama hit your CPU for 5 minutes with no output
minus-squareSeefra 1@lemmy.ziplinkfedilinkarrow-up6·2 months agoIsn’t that true for most models until someone destiles and quantises them so they can run on common hardware?
minus-squarefuckwit_mcbumcrumble@lemmy.dbzer0.comlinkfedilinkEnglisharrow-up8·2 months agoThis is the internet, we’re only allowed to be snarky here.
minus-squareGhoelian@lemmy.dbzer0.comlinkfedilinkarrow-up1·edit-22 months agoI mean yeah, but that doesn’t make the title any more true.
minus-squareCyberSeeker@discuss.tchncs.delinkfedilinkEnglisharrow-up1·2 months agoYes, but 20 billion parameters is too much for most GPUs, regardless of quantization. You would need at least 14GB, and even that’s unlikely without offloading major parts to the CPU and system RAM (which kills the token rate).
*if you have a laptop with 16gb of vram. Otherwise you’ll be watching ollama hit your CPU for 5 minutes with no output
Isn’t that true for most models until someone destiles and quantises them so they can run on common hardware?
This is the internet, we’re only allowed to be snarky here.
I mean yeah, but that doesn’t make the title any more true.
Yes, but 20 billion parameters is too much for most GPUs, regardless of quantization. You would need at least 14GB, and even that’s unlikely without offloading major parts to the CPU and system RAM (which kills the token rate).