ViSPLA: Visual Iterative Self-Prompting for Language-Guided 3D Affordance Learning | Read Paper on Bytez