ai-models/nunchaku-qwen-image-edit-2509
2
0
Fork 0
mirror of https://www.modelscope.cn/nunchaku-tech/nunchaku-qwen-image-edit-2509.git synced 2026-04-02 19:12:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Upload ./README.md to ModelScope hub

Browse Source
This commit is contained in:
Lmxyy1999
2025-11-16 02:24:23 +00:00
parent 10b3909f78
commit 829832625a
1 changed files with 63 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
README.md
View File

@ -34,10 +34,10 @@ This repository contains Nunchaku-quantized versions of [Qwen-Image-Edit-2509](h
## News ## 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](https://huggingface.co/lightx2v/Qwen-Image-Lightning/tree/main/Qwen-Image-Edit-2509). All models are available in the [`lightning-251115`](./lightning-251115) folder.
- [2025-09-25] 🔥 Release **4-bit [4/8-step lightning Qwen-Image-Edit](https://huggingface.co/lightx2v/Qwen-Image-Lightning)**! - [2025-09-25] 🔥 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! - [2025-09-24] 🚀 Release 4-bit SVDQuant quantized Qwen-Image-Edit-2509 model with rank 32 and 128!
## Model Details ## Model Details
### Model Description ### Model Description
@ -49,18 +49,67 @@ This repository contains Nunchaku-quantized versions of [Qwen-Image-Edit-2509](h
### Model Files ### Model Files
- [`svdq-int4_r32-qwen-image-edit-2509.safetensors`](./svdq-int4_r32-qwen-image-edit-2509.safetensors): SVDQuant INT4 (rank 32) Qwen-Image-Edit-2509 model. For users with non-Blackwell GPUs (pre-50-series). **Data Type**: `INT4` for non-Blackwell GPUs (pre-50-series), `NVFP4` for Blackwell GPUs (50-series).
- [`svdq-int4_r128-qwen-image-edit-2509.safetensors`](./svdq-int4_r128-qwen-image-edit-2509.safetensors): SVDQuant INT4 (rank 128) Qwen-Image-Edit-2509 model. For users with non-Blackwell GPUs (pre-50-series). It offers better quality than the rank 32 model, but it is slower. **Rank**: `r32` for faster inference, `r128` for better quality but slower inference.
- [`svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors): SVDQuant INT4 (rank 32) 4-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with non-Blackwell GPUs (pre-50-series).
- [`svdq-int4_r128-qwen-image-edit-2509-lightning`](./svdq-int4_r128-qwen-image-edit-2509-lightning): SVDQuant INT4 (rank 128) 4-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with non-Blackwell GPUs (pre-50-series).
- [`svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors): SVDQuant INT4 (rank 32) 8-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with non-Blackwell GPUs (pre-50-series).
- [`svdq-int4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-int4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors): SVDQuant INT4 (rank 128) 8-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with non-Blackwell GPUs (pre-50-series).
- [`svdq-fp4_r32-qwen-image-edit-2509.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509.safetensors): SVDQuant NVFP4 (rank 32) Qwen-Image-Edit-2509 model. For users with Blackwell GPUs (50-series).
- [`svdq-fp4_r128-qwen-image-edit-2509.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509.safetensors): SVDQuant NVFP4 (rank 128) Qwen-Image-Edit-2509 model. For users with Blackwell GPUs (50-series). It offers better quality than the rank 32 model, but it is slower. ### Base Models
- [`svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors): SVDQuant NVFP4 (rank 32) 4-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with Blackwell GPUs (50-series).
- [`svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-4steps.safetensors): SVDQuant NVFP4 (rank 128) 4-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with Blackwell GPUs (50-series). Standard inference speed models for general use
- [`svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors): SVDQuant NVFP4 (rank 32) 8-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with Blackwell GPUs (50-series).
- [`svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors): SVDQuant NVFP4 (rank 128) 8-step Qwen-Image-Edit-2509 model by fusing [Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) using LoRA strength = 1.0. For users with Blackwell GPUs (50-series).
| Data Type | Rank | Model Name | Comment |
|-----------|------|----------|---------|
| INT4 | r32 | [`svdq-int4_r32-qwen-image-edit-2509.safetensors`](./svdq-int4_r32-qwen-image-edit-2509.safetensors) | |
| | r128 | [`svdq-int4_r128-qwen-image-edit-2509.safetensors`](./svdq-int4_r128-qwen-image-edit-2509.safetensors) | |
| NVFP4 | r32 | [`svdq-fp4_r32-qwen-image-edit-2509.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509.safetensors) | |
| | r128 | [`svdq-fp4_r128-qwen-image-edit-2509.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509.safetensors) | |
### 4-Step Distilled Models
4-step distilled models fused with [Qwen-Image-Lightning-4steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) or [Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors) using LoRA strength = 1.0
| Data Type | Rank | Model Name | Comment |
|-----------|------|----------|---------|
| INT4 | r32 | [`svdq-int4_r32-qwen-image-edit-2509-lightning-4steps-251115.safetensors`](./lightning-251115/svdq-int4_r32-qwen-image-edit-2509-lightning-4steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors) |
| | r128 | [`svdq-int4_r128-qwen-image-edit-2509-lightning-4steps-251115.safetensors`](./lightning-251115/svdq-int4_r128-qwen-image-edit-2509-lightning-4steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors) |
| NVFP4 | r32 | [`svdq-fp4_r32-qwen-image-edit-2509-lightning-4steps-251115.safetensors`](./lightning-251115/svdq-fp4_r32-qwen-image-edit-2509-lightning-4steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors) |
| | r128 | [`svdq-fp4_r128-qwen-image-edit-2509-lightning-4steps-251115.safetensors`](./lightning-251115/svdq-fp4_r128-qwen-image-edit-2509-lightning-4steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-4steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors) |
| INT4 | r32 | [`svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors) | Fused with [Qwen-Image-Lightning-4steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) |
| | r128 | [`svdq-int4_r128-qwen-image-edit-2509-lightning`](./svdq-int4_r128-qwen-image-edit-2509-lightning) | Fused with [Qwen-Image-Lightning-4steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) |
| NVFP4 | r32 | [`svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-4steps.safetensors) | Fused with [Qwen-Image-Lightning-4steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) |
| | r128 | [`svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-4steps.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-4steps.safetensors) | Fused with [Qwen-Image-Lightning-4steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0-bf16.safetensors) |
### 8-Step Distilled Models
8-step distilled models fused with [Qwen-Image-Lightning-8steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) or [Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-8steps-V1.0-bf16.safetensors) using LoRA strength = 1.0
| Data Type | Rank | Model Name | Comment |
|-----------|------|----------|---------|
| INT4 | r32 | [`svdq-int4_r32-qwen-image-edit-2509-lightning-8steps-251115.safetensors`](./lightning-251115/svdq-int4_r32-qwen-image-edit-2509-lightning-8steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-8steps-V1.0-bf16.safetensors) |
| | r128 | [`svdq-int4_r128-qwen-image-edit-2509-lightning-8steps-251115.safetensors`](./lightning-251115/svdq-int4_r128-qwen-image-edit-2509-lightning-8steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-8steps-V1.0-bf16.safetensors) |
| NVFP4 | r32 | [`svdq-fp4_r32-qwen-image-edit-2509-lightning-8steps-251115.safetensors`](./lightning-251115/svdq-fp4_r32-qwen-image-edit-2509-lightning-8steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-8steps-V1.0-bf16.safetensors) |
| | r128 | [`svdq-fp4_r128-qwen-image-edit-2509-lightning-8steps-251115.safetensors`](./lightning-251115/svdq-fp4_r128-qwen-image-edit-2509-lightning-8steps-251115.safetensors) | 🔥 New: Fused with [Qwen-Image-Edit-2509-Lightning-8steps-V1.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Edit-2509/Qwen-Image-Edit-2509-Lightning-8steps-V1.0-bf16.safetensors) |
| INT4 | r32 | [`svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-int4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors) | Fused with [Qwen-Image-Lightning-8steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) |
| | r128 | [`svdq-int4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-int4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors) | Fused with [Qwen-Image-Lightning-8steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) |
| NVFP4 | r32 | [`svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-fp4_r32-qwen-image-edit-2509-lightningv2.0-8steps.safetensors) | Fused with [Qwen-Image-Lightning-8steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) |
| | r128 | [`svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors`](./svdq-fp4_r128-qwen-image-edit-2509-lightningv2.0-8steps.safetensors) | Fused with [Qwen-Image-Lightning-8steps-V2.0 LoRA](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0-bf16.safetensors) |
### Model Sources ### Model Sources
@ -68,7 +117,7 @@ This repository contains Nunchaku-quantized versions of [Qwen-Image-Edit-2509](h
- **Inference Engine:** [nunchaku](https://github.com/nunchaku-tech/nunchaku) - **Inference Engine:** [nunchaku](https://github.com/nunchaku-tech/nunchaku)
- **Quantization Library:** [deepcompressor](https://github.com/nunchaku-tech/deepcompressor) - **Quantization Library:** [deepcompressor](https://github.com/nunchaku-tech/deepcompressor)
- **Paper:** [SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models](http://arxiv.org/abs/2411.05007) - **Paper:** [SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models](http://arxiv.org/abs/2411.05007)
- **Demo:** [svdquant.mit.edu](https://svdquant.mit.edu) - **Demo:** [demo.nunchaku.tech](https://demo.nunchaku.tech)
## Usage ## Usage