Add Visual Jenga datasets: ClutteredParse, COCO, Full Scene Decom, and NYU with HTML visualizations f674cb4 Konpat Preechakul commited on Dec 9, 2025