Spaces:

Gustavosta

/

MagicPrompt-Dalle

like 38

Runtime error

In my experience, this model takes 3-4 seconds to produce prompts. The Stable diffusion magic prompt takes between 15 and 20 seconds. Why is that? This model produces shorter prompts, perhaps that is the reason. Or it is because this model was trained with a third of the steps (40,000 steps) than the stable diffusion model (150,000 steps). Any thoughts?