Conversation

Notices

1. kaia (kaia@brotka.st)'s status on Friday, 27-Oct-2023 06:53:13 JST kaia
   wait, could I get a Nvidia Tesla data center card and do Stable Diffusion??
   • LS (lain@lain.com)'s status on Friday, 27-Oct-2023 06:53:08 JST LS

     in reply to

     • Moon
     • iced quinnsmas :blobcatsanta:
     @kaia @icedquinn @Moon try an instance on runpod.io, a 4090 system costs around a dollar an hour.
   • Moon (moon@shitposter.club)'s status on Friday, 27-Oct-2023 06:53:09 JST Moon

     in reply to

     • iced quinnsmas :blobcatsanta:
     @kaia @icedquinn does amd have cuda now
   • kaia (kaia@brotka.st)'s status on Friday, 27-Oct-2023 06:53:09 JST kaia

     in reply to

     • Moon
     • iced quinnsmas :blobcatsanta:
     @Moon @icedquinn
     they have ROCm and PyTorch support and whatnot. I don't quite understand. it can generate Stable Diffusion, but it's slow. I would hope Nvidia would be way faster. but I won't fork over $2k to find out :sadcat:
   • kaia (kaia@brotka.st)'s status on Friday, 27-Oct-2023 06:53:11 JST kaia

     in reply to

     • iced quinnsmas :blobcatsanta:
     @icedquinn okay, I was told it's not worth it and my AMD GPU is way faster :asuka_hmm:
   • iced quinnsmas :blobcatsanta: (icedquinn@blob.cat)'s status on Friday, 27-Oct-2023 06:53:12 JST iced quinnsmas :blobcatsanta:

     in reply to
     @kaia teslas were used for a lot of ai training until newer cards came out
     so, probably. as long as it speaks cuda it will run pytorch
   • LS (lain@lain.com)'s status on Friday, 27-Oct-2023 06:56:35 JST LS

     in reply to

     • Moon
     • iced quinnsmas :blobcatsanta:
     @kaia @icedquinn @Moon i mostly generate locally but sometimes it's nice to be able to iterate super quickly
   • kaia (kaia@brotka.st)'s status on Friday, 27-Oct-2023 06:56:36 JST kaia

     in reply to

     • Moon
     • LS
     • iced quinnsmas :blobcatsanta:
     @lain @icedquinn @Moon
     oooh that sounds like much fun to explore. will try.