Instructions to use dx8152/Qwen-Image-Edit-2509-Relight with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use dx8152/Qwen-Image-Edit-2509-Relight with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline from diffusers.utils import load_image # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Qwen/Qwen-Image-Edit-2509", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("dx8152/Qwen-Image-Edit-2509-Relight") prompt = "Turn this cat into a dog" input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png") image = pipe(image=input_image, prompt=prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
wrong .json file the one on this is from the Fusion_lora
Hi There,
I think it may doesn't matter but the .json is the one from "Fusion_lora" workflow when I use I don't get the results as the dataset specially image 7 when going to sun lit to overcast. at same time all editors struggle with that but how you did to work?
你好,
我觉得这可能没什么关系,但是这个 .json 文件是来自“Fusion_lora”工作流程的,当我使用这个数据集时,特别是图 7,在从阳光照射到阴天时,我并没有得到正确的结果。所有编辑都遇到了这个问题,请问你是怎么做到的?
Hi There,
I think it may doesn't matter but the .json is the one from "Fusion_lora" workflow when I use I don't get the results as the dataset specially image 7 when going to sun lit to overcast. at same time all editors struggle with that but how you did to work?
你好,
我觉得这可能没什么关系,但是这个 .json 文件是来自“Fusion_lora”工作流程的,当我使用这个数据集时,特别是图 7,在从阳光照射到阴天时,我并没有得到正确的结果。所有编辑都遇到了这个问题,请问你是怎么做到的?
Sorry, I messed up some of the resources while organizing them. I've corrected the workflow.
Regarding Figure 7, you're correct. The model can't transition from strong sunlight to soft sunlight, so I reversed it and used it for training. This is the training dataset. Should I delete this image?
I just tested it, and even after training, LoRa can't erase shadows. This seems to be an irreversible process. I saw some feedback on YouTube about this issue, so I think it's best to delete this image.