thedeoxen/MOSS-Image-segmentation-controlnet

MOSS-Image-SFT — Segmentation Control LoRA

Short description

A segmentation Control LoRA for MOSS-Image-SFT. Feed a flat-color region map. The model follows the region boundaries, and the prompt decides what each region becomes.

This is a ControlNet-style edit LoRA, not the official MOSS color/depth concat adapter. Load it with a LoRA loader and pass one control image through Text Encode MOSS-Image Edit.

Selected checkpoint: training step 4500.

Files

• Weights: moss_image_sft_segmentation_control_lora.safetensors
• ComfyUI workflow: moss_image_segmentation_control.json
• Repo: thedeoxen/MOSS-Image-segmentation-controlnet

Examples

Each preview is one combined image, left to right: control, then three prompts.

Control → Results

Extended description

Follows flat color regions. Boundaries and the big masses stay, while the prompt fills them with a new subject and style. Appearance, materials, clothing and the world come from the text prompt.

The LoRA is rank 64 on linear layers and rank 16 on convolutions. It was trained as an edit adapter: the control image is the reference, and the prompt describes the picture to generate.

How to use

1. Load MOSS-Image-SFT: UNet moss_image_sft_int8_convrot.safetensors, text encoder qwen3.5_9b_w4a8_gptq.safetensors with type moss_image, VAE flux2-vae_small.safetensors.
2. Load this LoRA on the UNet at strength 1.0.
3. Build the control map and pass it as the single reference image in Text Encode MOSS-Image Edit.
4. Set edit format to ai-toolkit on both the positive and the negative encode. This LoRA was trained with kv cache on.
5. Sample with 30 steps, CFG 4, euler + simple.

No trigger word. The control image sets the structure. The prompt sets identity, materials, clothing and the world.

ComfyUI

The workflow in this repo loads a photo, builds the control map, encodes it with the prompt, and samples.

The included workflow uses the UniFormer semantic segmentor. The training maps are fully filled flat colors, with no large black unlabeled areas. The colors are region ids, not the colors you want in the result.

Also uses the built-in Text Encode MOSS-Image Edit node. Set the edit format to ai-toolkit.

Base model

Trained on and recommended with OpenMOSS-Team/MOSS-Image-SFT.

Typical settings from the included workflow: 30 steps, CFG 4, euler + simple, LoRA strength 1.0.

Negative prompt used in training samples:

nsfw, worst quality, bad anatomy, watermark, jpeg artifact

Example prompt

Full-body photograph of a circus ringmaster in scarlet tails, gold braid and a tall black hat, standing three-quarter view in front of a painted circus tent. One arm raised high, the other holding a slender black cane down at the side. Confetti, saturated color, entire body in frame.

What it does

• Follows flat color regions. Boundaries and the big masses stay, while the prompt fills them with a new subject and style.
• Lets the prompt define identity, materials, clothing and style.
• Works as a single-image control adapter for MOSS-Image-SFT.

Tips

• Flat, solid colors. Gradients and textures in the control map confuse the regions.
• Do not expect a region color to become that color in the picture. Say the material in the prompt.
• UniFormer maps match these weights better than a segmentor that leaves the background black.
• Lower the LoRA strength toward 0.7 if the control is too rigid. Strength 1.0 is the trained value.

Use cases

• Blocking a scene into regions, then restyling each mass from the prompt.
• Keeping a layout of people, furniture and background while changing the render.
• Color-script layouts where only the shapes are decided.