Shape Classification on ScanObjectNN
68.7Top-1 AccuracyHOLA-PointBERT
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| HOLA-PointBERT#Params=72.1M, FLOPs=84G, FPS=152, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 68.7 | — | — | 86.5 | 93.1 | |
| HOLA-PointBERT#Params=32.3M, FLOPs=29G, FPS=202, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 68.3 | — | — | 86.7 | 93 | |
| HOLA-PointBERT#Params=72.1M, FLOPs=84G, FPS=152, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 67.7 | — | — | 85.4 | 92.4 | |
| HOLA-PointBERT#Params=32.3M, FLOPs=29G, FPS=202, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 67.2 | — | — | 85.4 | 91.9 | |
| Uni3Dtraining shape source=Ensembled (no LVIS)2023.10 | 66.5 | — | — | 83.5 | 90.1 | |
| UNI3D-G#Params=1020.0M, FLOPs=1130G, FPS=34, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 66.5 | — | — | 83.5 | 90.1 | |
| RECON++-L#Params=658.9M, FLOPs=423G, FPS=34, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 65.4 | — | — | 84.1 | 89.7 | |
| Uni3D †training shape source=Ensembled, best achieved=true2023.10 | 65.3 | — | — | 85.5 | 92.7 | |
| UNI3D-G#Params=1020.0M, FLOPs=1130G, FPS=34, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 65.3 | — | — | 85.5 | 92.7 | |
| HOLA-PointBERT#Params=26.0M, FLOPs=7G, FPS=264, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 65.1 | — | — | 83.7 | 90.9 | |
| Uni3Dtraining shape source=Ensembled2023.10 | 63.9 | — | — | 84.9 | 91.7 | |
| RECON++-B#Params=201.5M, FLOPs=137G, FPS=90, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 63.6 | — | — | 80.2 | 90.6 | |
| HOLA-SparseConvPre-training regime=Ensembled (875,665 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 60.6 | — | — | 81.4 | 91 | |
| VIT-LENS-G#Params=1543.0M, FLOPs=806G, FPS=32, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 60.1 | — | — | 81 | 90.3 | |
| VIT-LENS-G#Params=2000.0M, FLOPs=1050G, FPS=27, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 59.8 | — | — | 79.3 | 87.7 | |
| HOLA-SparseConvPre-training regime=Ensembled no LVIS (829,460 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 59.3 | — | — | 79.8 | 89.4 | |
| MixCon3D-PointBERT#Params=30.9M, FLOPs=7G, FPS=153, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 58.6 | — | — | 80.3 | 89.2 | |
| TAMM-SparseConvPre-training regime=Ensembled (875,665 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 58.5 | — | — | 81.3 | 89.5 | |
| TAMM-SparseConvPre-training regime=ShapeNet (52,470 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 57.9 | — | — | 75.3 | 83.1 | |
| TAMM-SparseConvPre-training regime=Ensembled no LVIS (829,460 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 57.5 | — | — | 81.3 | 90 | |
| OpenShape-SparseConvtraining shape source=Ensembled2023.10 | 56.7 | — | — | 78.9 | 88.6 | |
| TAMM-PointBERT#Params=35.4M, FLOPs=29G, FPS=85, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 56.7 | — | — | 78.3 | 86.1 | |
| OpenShape-SparseConvPre-training regime=Ensembled (875,665 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 56.7 | — | — | 78.9 | 88.6 | |
| HOLA-SparseConvPre-training regime=ShapeNet (52,470 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 56.3 | — | — | 77.4 | 87.6 | |
| HOLA-PointBERT#Params=32.3M, FLOPs=29G, FPS=202, Training data=ShapeNet (52,470 triplets), Evaluation protocol=Zero-shot2026.05 | 55.9 | — | — | 77.4 | 87.7 | |
| TAMM-PointBERT#Params=35.4M, FLOPs=29G, FPS=85, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 55.7 | — | — | 80.7 | 88.9 | |
| OpenShape-SparseConvtraining shape source=Ensembled (no LVIS)2023.10 | 54.9 | — | — | 76.8 | 87 | |
| OpenShape-SparseConvPre-training regime=Ensembled no LVIS (829,460 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 54.9 | — | — | 76.8 | 87 | |
| TAMM-PointBERT#Params=35.4M, FLOPs=29G, FPS=85, Training data=ShapeNet (52,470 triplets), Evaluation protocol=Zero-shot2026.05 | 54.8 | — | — | 74.5 | 83.3 | |
| OpenShape-SparseConvPre-training regime=ShapeNet (52,470 triplets), Backbone=SparseConv (MinkowskiNet), Zero-shot=true2026.05 | 52.7 | — | — | 72.7 | 83.6 | |
| OpenShape-PointBERTtraining shape source=Ensembled2023.10 | 52.2 | — | — | 79.7 | 88.7 | |
| OpenShape-PointBERT#Params=32.3M, FLOPs=29G, FPS=86, Training data=Ensembled (875,665 triplets), Evaluation protocol=Zero-shot2026.05 | 52.2 | — | — | 79.7 | 88.7 | |
| ULIP-PointBERTtraining shape source=Ensembled2023.10 | 51.6 | — | — | 72.5 | 82.3 | |
| OpenShape-PointBERT#Params=5.1M, FLOPs=1G, FPS=240, Training data=ShapeNet (52,470 triplets), Evaluation protocol=Zero-shot2026.05 | 51.3 | — | — | 69.4 | 78.4 | |
| OpenShape-PointBERTtraining shape source=Ensembled (no LVIS)2023.10 | 47.2 | — | — | 72.4 | 84.7 | |
| OpenShape-PointBERT#Params=32.3M, FLOPs=29G, FPS=86, Training data=Ensembled no LVIS (829,460 triplets), Evaluation protocol=Zero-shot2026.05 | 47.2 | — | — | 72.4 | 84.7 | |
| ULIP-PointBERTtraining shape source=Ensembled (no LVIS)2023.10 | 46 | — | — | 66.1 | 76.4 | |
| DGCNN2023.03 | — | 73.6 | 78.1 | — | — | |
| GBNet2023.03 | — | 77.8 | 80.5 | — | — | |
| Point-BERT2023.03 | — | — | 83.1 | — | — | |
| Point-MAE2023.03 | — | — | 85.2 | — | — | |
| PointCNN2023.03 | — | 75.1 | 78.5 | — | — | |
| PointConTvoting strategy=false2023.03 | — | 86 | 88 | — | — | |
| PointConTvoting strategy=true2023.03 | — | 88.5 | 90.3 | — | — | |
| PointMLP2023.03 | — | 84.4 | 85.7 | — | — | |
| PointNet2023.03 | — | 63.4 | 68.2 | — | — | |
| PointNet++2023.03 | — | 75.4 | 77.9 | — | — | |
| PointNeXt2023.03 | — | 86.8 | 88.2 | — | — | |
| PRANet2023.03 | — | 79.1 | 82.1 | — | — | |
| SimpleView2023.03 | — | — | 80.5 | — | — | |
| SpiderCNN2023.03 | — | 69.8 | 73.7 | — | — |