Research
Qwen-Image-PE learns visual design for text-to-image prompt enhancement
Qwen-Image-PE uses supervised fine-tuning with caption-based reverse distillation and reinforcement learning with a groupwise design reward. The 397B variant improves over original requests by 7.81 points on Qwen-Image-Bench, and the 9B model is released on Hugging Face.
Every outlet that ran it 1 article
Hugging Face PapersResearchLead
Qwen-Image-PE: Learning Visual Design for Text-to-Image Prompt Enhancement