Recommended settings for ComfyUI.

#8
by razvanab - opened

There is support for Lumina-Image-2.0 in ComfyUI now. So, I would like to know more about the recommended settings, like the sampler name, for example.

The model car shows this:

height=1024,
width=1024,
guidance_scale=4.0,
num_inference_steps=50,
cfg_trunc_ratio=0.25,
cfg_normalization=True,

I don't know of any node in ComfyUI with these options.:
cfg_trunc_ratio and cfg_normalization

Alpha-VLLM org

@Nap Your setting looks great! You might also try reducing the sample steps to around 18 for faster results without compromising image quality.

Sign up or log in to comment