peteromallet/Qwen-Image-Edit-InStyle

1---2license: apache-2.03datasets:4- peteromallet/high-quality-midjouney-srefs5base_model:6- Qwen/Qwen-Image-Edit7tags:8- image9- editing10- lora11- style-transfer12- qwen13pipeline_tag: image-to-image14library_name: diffusers15---16 17# QwenEdit InStyle LoRA18 19## Model Description20 21**QwenEdit InStyle** is a LoRA fine-tune for QwenEdit that significantly improves its ability to generate images based on a style reference. While the base model has style transfer capabilities, it often misses the nuances of styles and can transplant unwanted details from the input image. This LoRA addresses these limitations to provide more accurate style-based image generation.22 23![samples.png](samples.png)24 25## How to Use26 27To get the best results, start your prompt with the phrase:28 29`Make an image in this style of `30 31And then describe what you want to generate.32 33For example:34`Make an image in this style of a serene mountain landscape at sunset.`35 36### use with diffusers37 38```39import torch40from diffusers import QwenImageEditPipeline41 42pipe = QwenImageEditPipeline.from_pretrained("Qwen/Qwen-Image-Edit", torch_dtype=torch.bfloat16)43pipe.to("cuda")44 45pipe.load_lora_weights("peteromallet/Qwen-Image-Edit-InStyle", weight_name="InStyle-0.5.safetensors")46```47 48### Strengths & Weaknesses49 50The model excels at:51- Capturing subtle style nuances from reference images52- Avoiding unwanted detail transplantation from source images53- Generating coherent images that match the intended style54- Strong prompt adherence55 56The model may struggle with:57- Extremely abstract or unconventional artistic styles58- Very specific technical details that conflict with the reference style59- Occasional anatomy issues60 61## Training Data62 63The QwenEdit InStyle LoRA was trained on a curated dataset of high-quality Midjourney style references. This dataset focuses on diverse artistic styles and provides clean style-content separation for better training.64 65You can find the public dataset used for training here:66[https://huggingface.co/datasets/peteromallet/high-quality-midjouney-srefs](https://huggingface.co/datasets/peteromallet/high-quality-midjouney-srefs)67 68## Links69 70- Model: [https://huggingface.co/peteromallet/Qwen-Image-Edit-InStyle](https://huggingface.co/peteromallet/Qwen-Image-Edit-InStyle)71- Dataset: [https://huggingface.co/datasets/peteromallet/high-quality-midjouney-srefs](https://huggingface.co/datasets/peteromallet/high-quality-midjouney-srefs)