Sign In

Neta Lumina Lightning LoRA

12

70

4

Updated: Nov 11, 2025

style

Verified:

SafeTensor

Type

LoRA

Stats

57

0

Reviews

Published

Nov 10, 2025

Base Model

Lumina

Usage Tips

Strength: 1

Hash

AutoV2
30162B8E4F

Neta Lumina Lightning LoRA

All cover images are raw output from the model, 1MP resolution, no upscale, no hands/faces inpainting fixes, even no negative prompt.


They said Lumina 2 is slow.

What about 2x faster 🚀.

This model is CFG distilled version of Neta Lumina.

About quality: Same as original model. I don't see decrease in quality, seems unnoticeable.

What is CFG distilled: In short: CFG distilled model learned positive + negative prompts + CFG all together during training. So it can output all results in one forward pass, no need to run second forward pass for negative prompt, thus 2x faster. At the cost of you can't control negative prompt and CFG scale. Unlike step distillation which is very aggressive, CFG distillation is simple and straightforward. So the quality penalty is unnoticeable.

Currently support:


How to use

Required settings to work:

  • LoRA strength 1

  • CFG scale 1 (This disables CFG, and negative prompt)

Recommended settings for better quality:

  • Timesteps shift 3 (from node ModelSamplingAuraFlow)

  • Steps > 20

  • Sampler: euler or euler_ancestral (euler_a). Note: most fancy samplers do not work.

Tips:

Use this ComfyUI plugin for more optimizations: https://civitai.com/articles/22251

  • 3x faster on older gpu (rtx 20xx and before) by running in fp16 mode.

  • 30% faster by using torch.compile.

With proper setup Lumina 2 can be "almost" as fast as SDXL. "almost" means 80%.

If you really need to remove something using negative prompt and still want to generate faster. You can split your KSampler to two. First part (20% steps), normal settings with CFG. Second part (rest 80% steps), load this LoRA and set CFG to 1.


FAQ:

Does this LoRA support low steps?

No. This is not a step distillation. Step distillation is too much for me, as well as for the model.


License 

This model is released under Apache License 2.0.

Additional terms: Selling or monetizing models that merged this LoRA is prohibited.


Update logs

v0.2: init version

  • mimics the effect of CFG 3.5. Good for details.

  • v0.2b: This version mimics the effect of CFG 5. Good for stability. Also smaller file size.


You have read this far, don't forget to leave a feedback in discussion. Don't write a review, Civitai review system is so hard to find and, most importantly, read...