Update README.md
Browse filesAWPortraitXL 1.1 has been fine-tuned with nearly 2,000 high-aesthetic-quality fashion photography images on top of version 1.0. This enhancement significantly improves composition aesthetics and output details. Compared to version 1.0, version 1.1 has a more precise response to different ethnicities, with more refined skin texture representations that cover various races. It boasts strong semantic understanding capabilities, providing multi-scene, high-quality, stable, and realistic styles in direct output. The model has enhanced recognition of oriental elements such as cheongsam and hanfu. Compared to other SDXL realistic models, it offers a higher aesthetic appeal for East Asian ethnicities. For more examples of model effects, please refer to the example prompts.
To achieve the best image quality, please follow these parameter settings:
Sampling Recommendations: DPM++ 2M Karras, Restart, Euler a, Step 30
CFG (Classifier-Free Guidance): Recommended value of 3
Resolution: 768x1152
Upscaling Algorithm: Lanczos, with a redraw scale of 0.2-0.3
Negative Prompt Keywords: (worst quality), (low quality), (normal quality), lowres, bad anatomy, ((bad hands, broken hands)), (many fingers), ((grayscale)), watermark, moles
Adetailer Settings:
Close-up: May be turned off or set to a redraw scale of 0.25-0.3
Upper Body: May be turned off or set to a redraw scale of 0.3-0.35
Full Body: Should be set to a redraw scale of 0.35-0.4
Model Usage Agreement (Please read carefully)
By downloading this model or using any online image generation service based on this model, you agree to comply with this agreement.
Individual Users: You may use this model for image generation (including commercial use) and for fusion models based on research and learning purposes. If you redistribute derivative models that fuse AWPortrait, you must open-source them and clearly indicate in the model description that AWPortrait was used, including disclosing the fusion weights. Derivative models may not be sold or used commercially (including but not limited to text-to-image, image-to-image, face-swapping, integration into any third-party software or services, etc.). If you violate this agreement and gain commercial income or cause any loss to the author, the author has the right to seek compensation.
Corporate Users (including but not limited to companies and profit-making organizations): If you wish to use this model (including but not limited to text-to-image, image-to-image, face-swapping, integration into any third-party software or services, etc.), you must purchase a commercial license for the model. To purchase a license, please visit Xiaohongshu. If you violate this agreement and gain commercial income or cause any loss to the author, the author has the right to seek compensation.
Disclaimer:
The content generated using this model service is maintained by you and used at your own discretion. You must comply with applicable laws, regulations, and the terms of this service. Any intellectual property issues arising from content generated by this service are your responsibility, and you are liable for the content generated. The author is not responsible for any losses caused. Except as otherwise provided by law, without the author's written consent, this model service may not be integrated into any third-party software or services, nor may it be disclosed to any third party directly or indirectly, explicitly or implicitly, regarding the connection between the generated content and this service. Non-individual users (including but not limited to companies and profit-making organizations) may not use the model for commercial purposes (including but not limited to text-to-image, image-to-image, face-swapping, etc.) without the author's written consent. If you violate this agreement and gain commercial income or cause any loss to the author, the author has the right to seek compensation.