☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to technology@hexbear.netEnglish · 16 days agoKimi Linear is a hybrid linear attention architecture that outperforms traditional full attention methods across various contextsgithub.comexternal-linkmessage-square3linkfedilinkarrow-up117arrow-down10cross-posted to: technology@lemmy.ziptechnology@lemmy.mltechnology@lemmygrad.ml
arrow-up117arrow-down1external-linkKimi Linear is a hybrid linear attention architecture that outperforms traditional full attention methods across various contextsgithub.com☆ Yσɠƚԋσʂ ☆@lemmygrad.ml to technology@hexbear.netEnglish · 16 days agomessage-square3linkfedilinkcross-posted to: technology@lemmy.ziptechnology@lemmy.mltechnology@lemmygrad.ml
minus-squarePM_ME_VINTAGE_30S [he/him]@lemmy.sdf.orglinkfedilinkEnglisharrow-up1·16 days agoEagerly awaiting a quantized version so one of my computers can run it. (Best I have access to right now has 32 GB of RAM.)
Eagerly awaiting a quantized version so one of my computers can run it. (Best I have access to right now has 32 GB of RAM.)