ComfyUI_CaptionThis

Describe a single image or all images in a directory using models such as Janus Pro, Florence2, or JoyCaption (testing), with a particular focus on building datasets for training LoRA.