---
title: Qwen-Edit-2509-Upscale-LoRA
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA.md"
repository: AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA
last_updated: 2026-06-17
license: gpl-3.0
pipeline_tag: image-to-image
tasks:
  - image-to-image
base_model:
  - Qwen/Qwen-Image-Edit-2509
base_model_relation: adapter
parameters: 18.4B
tensor_type:
  - BF16
library_name:
  - lora
  - safetensors
  - pytorch
frameworks:
  - pytorch
supports_inference: img2img
downloads: 653
stars: 5
tags:
  - lora
---

# Qwen-Edit-2509-Upscale-LoRA

> Qwen-Edit-2509-Upscale-LoRA - AI-ModelScope 在 ModelScope 开源的模型。Still work in progress, checkpoints are being uploaded on the go

AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA 是 ModelScope 魔搭社区上的 18.4B 参数image-to-image模型，采用 gpl-3.0 许可，基于 Qwen/Qwen-Image-Edit-2509 构建，并支持在线推理（img2img）。

- **Repository**: AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA
- **License**: gpl-3.0
- **Tasks**: image-to-image
- **Parameters**: 18.4B
- **Base model**: Qwen/Qwen-Image-Edit-2509
- **Online inference**: img2img
- **Tags**: lora
- **Downloads**: 653
- **Stars**: 5
- **Last updated**: 2026-06-17

Source: https://www.modelscope.cn/models/AI-ModelScope/Qwen-Edit-2509-Upscale-LoRA

---

![2025-11-08 03-09-06 (0-00-14-28)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/88-IXOJlZIVU5p6pkvXIL.jpeg)

**Still work in progress, checkpoints are being uploaded on the go**

**Current best:** LoRA A: qwen-edit-enhance_64-v3_000001000, LoRA B: qwen-edit-enhance_000004250

**V4 Candidates** LoRA A: [qwen-edit-enhance-v4-16_a_000003000](https://huggingface.co/vafipas663/Qwen-Edit-2509-Upscale-LoRA/blob/main/archive/qwen-edit-enhance-v4-16_a_000003000.safetensors)/1000, LoRA B: [qwen-edit-enhance-v4-64_a_000003750](https://huggingface.co/vafipas663/Qwen-Edit-2509-Upscale-LoRA/blob/main/archive/qwen-edit-enhance-v4-64_a_000003750.safetensors)/4750

LoRA A and LoRA B are meant to be wired one after another

Comfy workflow [is in the files](https://huggingface.co/vafipas663/Qwen-Edit-2509-Upscale-LoRA/blob/main/Qwen-Edit-Upscale.json)

Demo: 🤗 [Click here](https://huggingface.co/spaces/akhaliq/Qwen-Edit-2509-Upscale-LoRA)

Prompt:
Enhance image quality, followed by textual description of the scene. The more descriptive it is, the better the upscale effect will be

All images below are from 8 step Lighting LoRA.
* ModelSamplingAuraFlow is a must, shift must be kept **below** 0.3. With higher resolutions, such as image 3, you can set it as low as 0.02
* Samplers - LCM (best), Euler_Ancestral, then Euler
* Schedulers all work and give varying results in terms of smoothness


Trained on a filtered subset of Unsplash-Lite and UltraHR-100K at 2.4MP
* Style: photography
* Subjects include: landscapes, architecture, interiors, portraits, plants, vehicles, abstract photos, man-made objects, food
* Trained to recover from:
> * Low resolution up to 16x
> * Oversharpened images
> * Noise up to 50%
> * Gaussian blur radius up to 3px
> * JPEG artifacts with quality as low as 5%
> * Motion blur up to 64px
> * Pixelation up to 16x
> * Color bands up to 3 bits
> * Images after upscale models - up to 16x

Not trained on: 2d, illustrations, artworks, text, graphics, CGI


![2025-11-08 03-09-07 (0-00-02-28)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/Y4IO7XGmwnBQ_Q070yu3z.jpeg)

![2025-11-08 03-09-06 (0-00-29-02)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/rCbKjjLXJms8h0tQdjY8V.jpeg)

![2025-11-08 03-09-06 (0-00-35-15)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/ugT4mSI5l5pyuv4Xlqfin.jpeg)

![2025-11-08 03-09-06 (0-00-56-16)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/Pv4KW5sWTnO4dS3Am7fDj.jpeg)

![2025-11-08 03-09-06 (0-01-09-04)](https://cdn-uploads.huggingface.co/production/uploads/67a640e47d0e88d64d08bd13/2cgL2ebOvuYQ-hdoXRZ7b.jpeg)