这次的Lora模型是训练生成任意白色背景的单体物体，包含人物、动物、工业产品、电子产品、家具、玩具模型等，目的是能够根据需求利用该Lora能更加稳定的生成纯白背景的物体图片，方便我们进行后期处理。初期测试，该lora模型特别适配lyriel系列大模型，其他2.5d和3d风格base模型可能要增加相匹配的Prompt和调整图片尺寸。目前对一些2d基础大模型适配情况一般，这个可能是由于我训练的素材基本都是真实的3d物体图片。后期可增加对应的2d风格的图片进行优化训练，希望大家可以尝试用一下此Lora，如果能提出您的建议和实验结果可以分享出来，我这边进行改进，谢谢！ This LoRA model has been trained to generate individual objects with a pure white background, including people, animals, industrial products, electronic devices, furniture, toy models, and more. Its purpose is to provide a stable generation of object images with a white background, facilitating post-processing tasks according to specific requirements. In initial testing, this LoRA model has shown particularly good compatibility with the Lyriel series of large models. However, for other 2.5D and 3D style base models, it may be necessary to adjust the matching prompts and image sizes accordingly. Currently, the adaptation of certain 2D base models is average, possibly due to the fact that the training dataset primarily consisted of real 3D object images. To improve this, I plan to include optimized training using 2D style images in the future. I encourage you to try out this LoRA model, and if you have any suggestions or experimental results to share, I would be grateful to receive them for further improvement. Thank you!

Files with SHA256: 42761e81ddd78d63a107b7f92dbe2aafc2a1f5978593c25474f333d808f539d4