nunchaku-ai/nunchaku-qwen-image-edit-2509

<p align="center" style="border-radius: 10px"> <img src="https://huggingface.co/datasets/nunchaku-tech/cdn/resolve/main/nunchaku/assets/nunchaku_v2.png" width="30%" alt="Nunchaku Logo"/> </p>

<div align="center"> <a href=https://discord.gg/Wk6PnwX9Sm target="blank"><img src=https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fdiscord.com%2Fapi%2Finvites%2FWk6PnwX9Sm%3Fwithcounts%3Dtrue&query=%24.approximatemembercount&logo=discord&logoColor=white&label=Discord&color=green&suffix=%20total height=22px></a> <a href=https://huggingface.co/datasets/nunchaku-tech/cdn/resolve/main/nunchaku/assets/wechat.jpg target="_blank"><img src=https://img.shields.io/badge/WeChat-07C160?logo=wechat&logoColor=white height=22px></a> </div>

Model Card for nunchaku-qwen-image-edit-2509

This repository contains Nunchaku-quantized versions of Qwen-Image-Edit-2509, an image-editing model based on Qwen-Image, advances in complex text rendering. It is optimized for efficient inference while maintaining minimal loss in performance.

News

• —[2025-11-15] 🚀 Release new quantized qwen-image-edit-2509 4/8-step lightning models, fused with lightx2v Qwen-Image-Edit-2509 lightning lora. All models are available in the `lightning-251115` folder.
• —[2025-09-25] 🔥 Release 4-bit [4/8-step lightning Qwen-Image-Edit](https://huggingface.co/lightx2v/Qwen-Image-Lightning)!
• —[2025-09-24] 🚀 Release 4-bit SVDQuant quantized Qwen-Image-Edit-2509 model with rank 32 and 128!

Model Details

Model Description

• —Developed by: Nunchaku Team
• —Model type: image-to-image
• —License: apache-2.0
• —Quantized from model: Qwen-Image-Edit-2509

Model Files

Data Type: INT4 for non-Blackwell GPUs (pre-50-series), NVFP4 for Blackwell GPUs (50-series). Rank: r32 for faster inference, r128 for better quality but slower inference.

Base Models

Standard inference speed models for general use

4-Step Distilled Models

4-step distilled models fused with Qwen-Image-Lightning-4steps-V2.0 LoRA or Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA using LoRA strength = 1.0

8-Step Distilled Models

8-step distilled models fused with Qwen-Image-Lightning-8steps-V2.0 LoRA or Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA using LoRA strength = 1.0

Model Sources

• —Inference Engine: nunchaku
• —Quantization Library: deepcompressor
• —Paper: SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models
• —Demo: demo.nunchaku.tech

Usage

• —Diffusers Usage: See qwen-image-edit-2509.py. Check this tutorial for more advanced usage.
• —ComfyUI Usage: See nunchaku-qwen-image-edit-2509.json.

Performance

Citation

@inproceedings{
  li2024svdquant,
  title={SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models},
  author={Li*, Muyang and Lin*, Yujun and Zhang*, Zhekai and Cai, Tianle and Li, Xiuyu and Guo, Junxian and Xie, Enze and Meng, Chenlin and Zhu, Jun-Yan and Han, Song},
  booktitle={The Thirteenth International Conference on Learning Representations},
  year={2025}
}