造梦实验室-日系写真-QWEN

EN：

[Large Language Model/Checkpoint] Qwen Image 20B v1

[Model Weight] 0.8~1 (For single LoRA, 1 is recommended)

[Trigger Word] rxxz

[Sampler Method] Euler

[CFG Scale] 3.5

[Sampling Steps] 30

[About Training] This model is trained using the Qwen Image 20B v1 base model.

[About Usage] This model adheres to the official Qwen license.

[Training Acknowledgements] This model's training received technical support from the ModelScope community.

ZH：

DreamLab-rxxz-qwen

造梦实验室-日系写真-QWEN

【大模型/Checkpoint】Qwen Image 20B v1

【模型权重】0.8~1（单lora建议直接上1）

【触发词】rxxz

【采样方法/Sampler Method】Euler

【提示词引导系数/CFG Scale】3.5

【采样步数/Sampling Steps】30

【关于训练】本模型底模使用Qwen Image 20B v1训练

【关于使用】本模型遵循Qwen官方许可证

【训练致谢】本模型训练获得魔搭社区技术支持

【推荐提示】

rxxz, photo of a young Japanese woman with short, light brown hair, wearing a denim jacket, standing in front of a wall covered in ivy.

rxxz, photo of a young Japanese man with glasses, wearing a simple sweatshirt, sitting on a wooden pier by a calm lake.

rxxz, photo of a single, empty wooden chair placed in a sunlit, abandoned room with peeling paint, dust motes visible in the sunbeams.