TurboClear: One-Step Object-Effect Removal via Region-Calibrated Distribution Matching and Fusion
Paper • 2608.01288 • Published
How to use JGuo666/TurboClear with Diffusers:
pip install -U diffusers transformers accelerate
import torch
from diffusers import DiffusionPipeline
from diffusers.utils import load_image
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("JGuo666/TurboClear", dtype=torch.bfloat16, device_map="cuda")
prompt = "Turn this cat into a dog"
input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png")
image = pipe(image=input_image, prompt=prompt).images[0]This repository contains the first TurboClear checkpoint release used by the public inference code.
TurboClear: One-Step Object-Effect Removal via Region-Calibrated Distribution Matching and Fusion. arXiv:2608.01288
sdxl/state_dict.pth: one-step DMD student generator state.fusion/fusion_module.pth: learnable spatial fusion head trained after the
generator.The SDXL/ObjectClear base model is not duplicated here. Check its license and download it separately before inference. OBER data and evaluation images are not redistributed.
Download the files, then pass the local sdxl directory as WEIGHT_PATH and
fusion/fusion_module.pth as FUSION_MODULE_PATH to
inference/inference_turboclear.sh.