Qwen-Image-Edit-2509-Light_restoration

This model is trained (code-free!) on ModelScope. Thanks to ModelScope team for providing the training infra:

https://www.modelscope.cn/aigc/modelTraining

---

/庆祝 Version 25.11.25 V2.0 Update

Fixed prompt words (no need to modify):

"移除光影,使用柔和光线（无明显光斑和阴影）对图片进行重新照明"

Online running link： https://www.runninghub.ai/post/1986847676127973378?inviteCode=rh-v1331

This is a user guide:

YouTube：https://youtu.be/SgCo_sDUquo

Blibili：https://www.bilibili.com/video/BV1neU9BzEEG/

---

To avoid the recurring question, "The photo looks better before removing the lighting, what's the point of this garbage LoRa?"

Let me preface this by saying that if you've ever trained a lighting LoRa dataset, you know how difficult it is to remove light from a person's face, but training requires a pair of images (one with lighting and one without).

For example, the previous approach to creating datasets was to first find a photo without lighting, and then use the lighting LoRa dataset to add lighting.

This approach has two drawbacks: 1. Photos without lighting are relatively hard to find. 2. The lighting added by the AI ​​never looks as natural as in a real photograph, and it requires a lot of trial and error.

Now, using this LoRa dataset, you can directly find any model image with the lighting you want to achieve, and use this LoRa to remove the lighting. The resulting lighting LoRa dataset is significantly better than previous datasets.

---

Hugging Face app

Online running link： https://www.runninghub.ai/post/1986847676127973378?inviteCode=rh-v1331

This is a user guide:

YouTube：https://youtu.be/ReKNKBQQky4

Blibili：https://www.bilibili.com/video/BV1Tc11BJETN/

Fixed prompt words (no need to modify):

" 移除光影,使用柔和光线对图片进行重新照明"

---

Instructions: Download the lora file to the models/loras folder.

You also need this lora and use them together: https://huggingface.co/lightx2v/Qwen-Image-Lightning/tree/main

For communication/cooperation, you can join the discord group to communicate： https://discord.gg/yVAVa43mWk

If these resources are helpful to you, or if you use them for business purposes, please buy me a coffee. Thank you for supporting original content! PayPal: Daniel8152