# H3 Sharpness LMS LoRa Successes? — retained record

> Retained reddit record from reddit:comfyui.

Canonical: https://highsignal.app/data/records/7a2e8060b9efe5ed

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:comfyui)  published 2026-09-21 08:40 ingested 2026-09-22 02:40

[original source ↗](https://www.reddit.com/r/comfyui/comments/1wm73j7/h3_sharpness_lms_lora_successes/)

## retained text



```
Has anyone had luck with the new "Sharpness" LMS LoRA Minimax workflow?  

I tried a dozen ways to Sunday, obediently follow instructions, tweaking it, feeding it different sized input videos, and all it seems to do it little more than a typical sharpen, no AI kicking in to hallucinate detail or anything.

Anyone got the trick to getting it to work?  I don't want to just discard with it, but maybe I am taken in by marketing puffery.  The images sure look convincing over on https://huggingface.co/Alissonerdx/Minimax-H3-ComfyUI.

Anyone gotten it to work right and produce actual advantage over other methods?  Thx,
```



document fetched 2026-09-22 00:41 · canonical https://reddit.com/r/comfyui/comments/1wm73j7/h3_sharpness_lms_lora_successes
