A small concept slider for MOSS Image that makes the hips and glutes of an adult figure fuller as you raise the strength. One file, no trigger word. It works best on back and three-quarter views where the lower body is fully in frame. Strength 0 leaves the picture unchanged. Read the limits below before you download.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎯 What it does (honest summary)
Same prompt and seed, MOSS Image SFT, 50 steps, with and without the LoRA. In our tests of 8 prompts (back, side and three-quarter views, different settings and skin tones) the hips and glutes got fuller as strength went up from 1.5 to 3.0 in most of them. The effect was small in two of the eight: a plain studio back view and a plain-wall back view. At strength 1.5 about 5 of 16 test pictures also changed things you did not ask for (hair, pose, shoes, background). We tested positive strengths only, not negative.
🖼️ Base vs LoRA. Same prompt and seed. Left: MOSS Image without the LoRA. Right: with this LoRA at the strength shown (2.0 or 2.5, see the file name in the gallery post).



🚀 Quick start
Base model: MOSS Image (trained and tested on MOSS Image SFT, int8)
Trigger word: none
Strength: start at 1.5 to 2.0. Up to 2.5 gives a stronger change. 3.0 is the maximum we tested
Settings used for the examples: SFT, 50 steps, CFG 3, euler / sgm_uniform, shift 5, 896x1344
Prompt tip: keep the whole lower body in frame, and describe the view (from behind, side profile)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚠️ Limits you should know
Above 2.5 the change can be very strong, and the pose, hair, camera angle or background can shift from the base picture (we saw this in some prompts at 3.0, and in one beach walking scene at every strength)
Small effect on plain studio or plain-wall back views in our tests
Tested on 8 prompts, mostly 1 seed per strength above 1.5, all adults. We did not test other styles, groups, or other base models
Trained once with one recipe. We did not run it multiple times to measure how much results vary between training runs
✨ Features
One small safetensors file, about 71 MB, no trigger word
Slider behaviour: strength 0 leaves the picture unchanged
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🧪 How it was made
Trained locally with ai-toolkit (concept slider) for 300 steps on adult pictures that I generated myself with MOSS Image. No photos of real people were used. All example images are AI-generated adults.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⭐ Found this useful?
Like if it saved you time
Comment your results — I read every one
Follow for new ComfyUI workflows, all tested on 16 GB VRAM
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚖️ Attribution and licence
Base model: MOSS-Image by OpenMOSS-Team, weights released under Apache-2.0 (model card). This LoRA is a modification trained on top of it
Trained with ai-toolkit by Ostris
Description
First release. Strength 1.5 to 2.0 recommended, up to 2.5 for a stronger change (3.0 is the maximum tested). Tested on MOSS Image SFT, 50 steps, CFG 3, euler / sgm_uniform, shift 5, 896x1344. No trigger word.







