qwen-image-2.1-face-lora-v1

Note: All model files below must be downloaded via the LM Downloader client. If you do not have the client, Download >>>

Qwen-Image-2.1 面容（亚洲人脸强化） LoRAThis LoRA is fine-tuned specifically for the Qwen-Image-2.1 base model to optimize Asian facial generation, mitigating common artifacts such as distorted facial features and broken faces from the base model, while enhancing skin texture and facial details for portraits. The LoRA is a standalone weight and cannot run independently; it must be used together with the Qwen-Image-2.1 base model. The recommended LoRA scale is 0.7–0.85. Higher scale values may lead to facial homogenization and overfitting. The workflow supports realistic portrait creation, capable of multi-person scenes, preserving environmental and object textures, and rendering full-sharp photographic images without background blur.