How to use from the
Use from the
nobg library
pip install nobg
# Option 1: use via the predict method

from nobg import AutoModel, AutoProcessor

model = AutoModel.from_pretrained("1038lab/MultiMatte").eval()
processor = AutoProcessor.from_pretrained("1038lab/MultiMatte")

cutout = model.predict(processor, "image.jpg", "prompt")
# Option 2: use the model and processor directly

import torch
from loadimg import load_img
from nobg import AutoModel, AutoProcessor

model = AutoModel.from_pretrained("1038lab/MultiMatte").eval()
processor = AutoProcessor.from_pretrained("1038lab/MultiMatte")

image = load_img("image.jpg").convert("RGB")
inputs = processor(image, return_tensors="pt")

with torch.no_grad():
    outputs = model(pixel_values=inputs["pixel_values"])

alpha = processor.post_process_alpha_matting(outputs, target_sizes=[(image.height, image.width)])[0]
processor.cutout(image, alpha).save("output.png")

MultiMatte

Promptable SAM 3 Alpha Matting β€” Packaged for ComfyUI-RMBG.

This repository hosts the weights of MultiMatte (Apache-2.0) for seamless integration with ComfyUI-RMBG.

MultiMatte is a promptable alpha matting model built on SAM 3 that outputs continuous opacity mattes for hair, glass, and fine edges from optional text prompts.

ComfyUI Usage

The MultiMatte (RMBG) node in ComfyUI-RMBG automatically downloads these files into:

ComfyUI/models/RMBG/MultiMatte/
β”œβ”€β”€ config.json
β”œβ”€β”€ model.safetensors
β”œβ”€β”€ processor_config.json
β”œβ”€β”€ tokenizer.json
└── tokenizer_config.json

License & Attribution

Downloads last month
-
Safetensors
Model size
0.8B params
Tensor type
F32
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support