From: carlf@panix.com   
      
   On 2024-12-02, Tristan Miller  wrote:   
      
   > I'm afraid I was reporting only half-remembered results from the last   
   > time I looked into the question, which would have been months ago.  I   
   > did a quick web search just now and came up with a couple queries from   
   > the Kohya's GUI GitHub project that roughly accord with my recollection:   
   >   
   > https://github.com/bmaltais/kohya_ss/discussions/679   
   > https://github.com/bmaltais/kohya_ss/issues/2632   
   >   
   > The first of these claims CPU training time of about 40×, and the second   
   > claims 4× for both time and memory.  They both refer to Stable Diffusion.   
   >   
   > There's also this LoRA finetuning guide for LLaMA that provides detailed   
   > CPU time and memory metrics for various models: https://rentry.org/cpu-lora   
      
   Thank you!   
   --   
   Carl Fink                                                  carl@finknetwork.com   
   https://reasonablyliterate.com                           https://nitpicking.com   
   If you want to make a point, somebody will take the point and stab you with it.   
   	-Kenne Estes   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)