Found A LoRA That Gives A Nicer Anime Look On Qwen-Image. Name Is Raena-Qwen-Image. License Is Apache-2.0.
What It Is
- Finetune for Qwen-Image, made for anime only
- Lines look clean, colors pop more, faces usually come out better
- Trained on ~500 hand-picked images with natural captions
How To Use
- Start prompt with: “Anime illustration of …” then your subject
- Set LoRA weight around 0.8 – 1.0 (adjust a little if needed)
- Works fine with Lightning LoRA if you want faster runs
ComfyUI Tip
- Sample images on the model page have metadata
- Download one and drag into ComfyUI to load a working graph
Download
Direct file (raena_qwen_image_lora_v0.1.safetensors):
Quick Notes
- Abstract / artsy styles: results can vary.
- Hard poses: hands or limbs may slip. Try LoRA weight 0.7–0.9 or nudge the prompt.
- Support link (Ko-Fi) is listed on the model page.