Leo@lemmy.linuxuserspace.showM to Linux and Tech News@lemmy.linuxuserspace.showEnglish · 3 days agoMozilla launches Thunderbolt AI client with focus on self-hosted infrastructurearstechnica.comexternal-linkmessage-square4linkfedilinkarrow-up18arrow-down13
arrow-up15arrow-down1external-linkMozilla launches Thunderbolt AI client with focus on self-hosted infrastructurearstechnica.comLeo@lemmy.linuxuserspace.showM to Linux and Tech News@lemmy.linuxuserspace.showEnglish · 3 days agomessage-square4linkfedilink
minus-squareBazell@lemmy.ziplinkfedilinkEnglisharrow-up1·2 days agoSelf hosted? Not sure that this will be a well working idea assuming how much RAM people have and will have in their systems nowadays.
minus-squareNotSteve_@lemmy.calinkfedilinkEnglisharrow-up3·2 days agoYou can actually run simple purpose-built models with very little RAM. I self host immich in a 4GB proxmox VM and it’s able to run its built in image subject recognition AI model without even breaking 50% of that 4GB of RAM
Self hosted? Not sure that this will be a well working idea assuming how much RAM people have and will have in their systems nowadays.
You can actually run simple purpose-built models with very little RAM. I self host immich in a 4GB proxmox VM and it’s able to run its built in image subject recognition AI model without even breaking 50% of that 4GB of RAM