Flying Squid@lemmy.world to Technology@lemmy.worldEnglish · 9 months agoYour AI Girlfriend Is a Data-Harvesting Horror Showgizmodo.comexternal-linkmessage-square96fedilinkarrow-up1494arrow-down113
arrow-up1481arrow-down1external-linkYour AI Girlfriend Is a Data-Harvesting Horror Showgizmodo.comFlying Squid@lemmy.world to Technology@lemmy.worldEnglish · 9 months agomessage-square96fedilink
minus-squareexu@feditown.comlinkfedilinkEnglisharrow-up4·9 months agoYou’ll want to use a quantised model on your GPU. You could also use the CPU and offload some parts to the GPU with llama.cpp (an option in oobabooga). Llama.cpp models are in the GGUF format.
You’ll want to use a quantised model on your GPU. You could also use the CPU and offload some parts to the GPU with llama.cpp (an option in oobabooga). Llama.cpp models are in the GGUF format.