# MOSS Image Style Training — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/ac7fc6e1583f690e

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


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-10-20 15:44 ingested 2026-10-21 02:38

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1xavlnq/moss_image_style_training/)

## retained text



```
Did anyone had any luck with training a complex style with MOSS? It resembles the style, but it's definitely something else, not even close. Tried to train 9 times with different caption style, different training parameters, etc, but the overall result is not it. I've also tried to up the learning rate x5 of the usual recommended number, but it just results in body-horror anatomy gens, and the style remains the same as before on lower LR. Any tips/help? 
```



document fetched 2026-10-21 00:30 · canonical https://reddit.com/r/StableDiffusion/comments/1xavlnq/moss_image_style_training
