This model was trained on synthetic data.
The training images were generated using
https://huggingface.co/lichorosario/lobato-qwen-image-lora (LoRA trained with qwen-image) but using qwen-image-2512 for inference.
Download them in the Files & versions tab.
