Yixi Style — MOSS-Image (r64)
A small personal mix, shared purely for reference and experimentation. This is not a trained model and not an original work of mine: it is simply a weight merge of several publicly available LoRAs, put together for my own use. All credit belongs to the original creators — if you enjoy the results, please visit and support their pages.
If any original author feels this should not be here, just let me know and I will take it down immediately, no questions asked.
Trigger word: yixi style
Base model: works on both MOSS-Image SFT and the base model. Strength 1.0 mirrors the mix I use; feel free to lower it to 0.7–0.9 if it starts to overpower your prompt. Results will vary with your own settings, so please treat the examples as a rough reference only.
Recommended acceleration LoRA (used for the example images):
https://civarchive.com/models/3006599/moss-image-turbo-loras?modelVersionId=3420002
Rank: 64, bf16 — a lighter file with slightly softer fine detail.
No dataset was used, since nothing was trained here. I make no claim of ownership over this file or the styles it contains.










