3D Semantic Segmentation on DAPS-3D 1.0 (test)
89.08mIoUHilDA
Evaluation results
| Method | Links | |
|---|---|---|
| HilDAFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 89.08 | |
| CleverDFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 87.95 | |
| HilDA†Fine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 87.18 | |
| LiMoEFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 86.68 | |
| LiMAFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 86.63 | |
| SuperFlowFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 86.21 | |
| HilDAFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 85.93 | |
| SealFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 85.9 | |
| ScaLRFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 85.58 | |
| SLidRFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 85.4 | |
| HilDA†Fine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 84.26 | |
| LiMoEFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 83.24 | |
| LiMAFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 83.11 | |
| CleverDFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 83.06 | |
| SuperFlowFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 82.43 | |
| ScaLRFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 81.92 | |
| SealFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 81.88 | |
| SLidRFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 81 | |
| RandomFine-tuning ratio=100%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 79.38 | |
| RandomFine-tuning ratio=50%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 74.32 |