Model Card for nunchaku-qwen-image-edit

This repository contains Nunchaku-quantized versions of Qwen-Image-Edit, 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.

No recent news. Stay tuned for updates!

Model Details

Model Description

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

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

Data Type	Rank	Model Name	Comment
INT4	r32	svdq-int4_r32-qwen-image-edit.safetensors
	r128	svdq-int4_r128-qwen-image-edit.safetensors
NVFP4	r32	svdq-fp4_r32-qwen-image-edit.safetensors
	r128	svdq-fp4_r128-qwen-image-edit.safetensors

4-Step Distilled Models

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

Data Type	Rank	Model Name	Comment
INT4	r32	svdq-int4_r32-qwen-image-edit-lightningv1.0-4steps.safetensors	Fused with Qwen-Image-Edit-Lightning-4steps-V1.0 LoRA
	r128	svdq-int4_r128-qwen-image-edit-lightningv1.0-4steps.safetensors	Fused with Qwen-Image-Edit-Lightning-4steps-V1.0 LoRA. Better quality, slower inference
NVFP4	r32	svdq-fp4_r32-qwen-image-edit-lightningv1.0-4steps.safetensors	Fused with Qwen-Image-Edit-Lightning-4steps-V1.0 LoRA
	r128	svdq-fp4_r128-qwen-image-edit-lightningv1.0-4steps.safetensors	Fused with Qwen-Image-Edit-Lightning-4steps-V1.0 LoRA. Better quality, slower inference

8-Step Distilled Models

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

Data Type	Rank	Model Name	Comment
INT4	r32	svdq-int4_r32-qwen-image-edit-lightningv1.0-8steps.safetensors	Fused with Qwen-Image-Edit-Lightning-8steps-V1.0 LoRA
	r128	svdq-int4_r128-qwen-image-edit-lightningv1.0-8steps.safetensors	Fused with Qwen-Image-Edit-Lightning-8steps-V1.0 LoRA. Better quality, slower inference
NVFP4	r32	svdq-fp4_r32-qwen-image-edit-lightningv1.0-8steps.safetensors	Fused with Qwen-Image-Edit-Lightning-8steps-V1.0 LoRA
	r128	svdq-fp4_r128-qwen-image-edit-lightningv1.0-8steps.safetensors	Fused with Qwen-Image-Edit-Lightning-8steps-V1.0 LoRA. Better quality, slower inference

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.py and qwen-image-edit-lightning.py. Check this tutorial for more advanced usage.
• ComfyUI Usage: See nunchaku-qwen-image-edit.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}
}