feat: add Ming-Image architecture & Prompt Enhancer and update upstream deps

feat: add Qwen-Image-2.1 prompt enhancer

refactor(injectors): consolidate boogu and qwen-2.1 into reference_image with flexible max_images

Add Qwen-Image-2.1 architecture model.

Add SenseNova-U1.5 architecture model.

Upgrade to Gradio 6