Conversation

Notices

1. kaia (kaia@brotka.st)'s status on Friday, 09-Jun-2023 22:27:54 JST kaia
   don't have enough RAM for Tensorflow on the laptop :sadcat:
   • LS (lain@lain.com)'s status on Friday, 09-Jun-2023 22:27:51 JST LS

     in reply to

     • mangeurdenuage
     @kaia @mangeurdenuage i do half of the stuff locally, the other half on runpod.io. runpod seems cheaper to me than actually buying a top end card, at least if you don't use it all the time.
   • mangeurdenuage (mangeurdenuage@shitposter.club)'s status on Friday, 09-Jun-2023 22:27:53 JST mangeurdenuage

     in reply to
     @kaia shitty software, shitty hardware requirement.
   • kaia (kaia@brotka.st)'s status on Friday, 09-Jun-2023 22:27:53 JST kaia

     in reply to

     • mangeurdenuage
     @mangeurdenuage
     thonking hard about buying Nvidia GPU machine to do more AI stuff :aaawoo:
   • mangeurdenuage (mangeurdenuage@shitposter.club)'s status on Friday, 09-Jun-2023 23:56:21 JST mangeurdenuage

     in reply to

     • Elicin
     @Elicin @kaia
     >Linux
     :gnu_interjection:
     
     Modular gpus should have been the defacto since forever, same goes for cpus, which are not a thing on most laptops these days.
     Ideally gpu, and gpu ram should be removable just like cpus/ram are on desktops.
     Heavens Feel likes this.
   • Elicin (elicin@freespeechextremist.com)'s status on Friday, 09-Jun-2023 23:56:25 JST Elicin

     in reply to

     • mangeurdenuage
     @mangeurdenuage does make a good point though, having a NVIDIA GPU to swap out on a Framework laptop might actually be a viable use for running one on Linux @kaia.
   • kaia (kaia@brotka.st)'s status on Friday, 09-Jun-2023 23:56:26 JST kaia

     in reply to

     • mangeurdenuage
     • LS
     @lain
     thank you, I'll check :kaia_look: @mangeurdenuage