Tencent's latest upgraded Hunyuan text-to-image large model (Hunyuan DiT) version 1.2 is compatible with Onijza, LoRA, ControlNet, and Kohya plugins in the Diffusers library, significantly lowering the threshold for training personalized image generation models. This upgraded version greatly enhances image quality and composition, and the Hunyuan Captioner model, which supports bilingual descriptions in Chinese and English, understands Chinese semantics more accurately, generating structured, complete, and accurate image descriptions. The open sourcing of the Hunyuan Captioner model allows.