TriCoLo: Trimodal Contrastive Loss for Text To Shape Retrieval | ComputerVisionFoundation Videos | Podwise