An updated version of this LORA, uses existing datasets but finetuned in kohya scripts, runs in local machine under FP8 training mode to cut down training time even further from taking 24+ hours to just 6+ hours.
An updated version of this LORA, uses existing datasets but finetuned in kohya scripts, runs in local machine under FP8 training mode to cut down training time even further from taking 24+ hours to just 6+ hours.