Published:
Activate LoRA (0 = off and ignore LoRA settings below, 1 or higher = on)
0
LoRA model (ALWAYS check LoRA description for best practices)
773250965756399630
Images per run (batch size)
4
Prompt
A white Husky dog sits calmly amidst chaos, its piercing blue eyes gazing innocently into the camera. The once-tidy living room now lies in disarray: vases shattered, pillows tossed aside, and coffee table scattered with books. Yet, our furry culprit appears oblivious to the destruction, sporting a 'who me?' expression that screams innocence. Framed by the destroyed decor, the dog's calm demeanor is a stark contrast to the mayhem it seems to have inspired.
Guidance (2 to 5 are safe values, below 2 might help with art styles, just avoid 1.0)
3.5
Sampler (avoid using sde's, ddpm and ancestrals)
heun
Scheduler (karras, ddim and exponential NOT recommended)
normal
Max shift (introduces more noise, lower values give less hallucinations, higher more details)
1.15
Base shift (has no effect in square images)
0.5