Conversation

Notices

1. kaia (kaia@brotka.st)'s status on Wednesday, 01-Nov-2023 04:41:32 JST kaia
   the 90 second prompt for one picture in StableDiffusion is down to 6.3 seconds :aqua_crazed:
   • LS likes this.
   • LS (lain@lain.com)'s status on Wednesday, 01-Nov-2023 04:42:43 JST LS

     in reply to
     @kaia i hope it didn't give you too much power...
   • kaia (kaia@brotka.st)'s status on Wednesday, 01-Nov-2023 04:50:22 JST kaia

     in reply to

     • LS
     @lain with great power comes great responsibility to generate the best possible tomboys.
     
     but without joking, this is alleviating so much pain, e.g. ComfyUI just installed the CUDA stuff and started on first try without any weird error, 15 required command line arguments for "low RAM" and whatnot :02smile:
     LS likes this.
   • LS (lain@lain.com)'s status on Wednesday, 01-Nov-2023 04:51:46 JST LS

     in reply to
     @kaia i have an amd chip in my rog ally and i thought 'hey let's see how that compares to my mac for SD' and i couldn't even get it to generate anything at all
   • LS (lain@lain.com)'s status on Wednesday, 01-Nov-2023 04:52:56 JST LS

     in reply to

     • LS
     @kaia although amd recently bought nod ai, they are experts at making AI stuff run cross platform, so at least they are taking things seriously.