3D Semantic Segmentation on SemiKITTI 1.0 (test)
59.02mIoUHilDA†
Evaluation results
| Method | Links | |
|---|---|---|
| HilDA†Fine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 59.02 | |
| HilDAFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 58.22 | |
| CleverDFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 56.7 | |
| LiMoEFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 55.96 | |
| LiMAFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 55.13 | |
| SuperFlowFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 54 | |
| SealFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 52.77 | |
| SLidRFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 50.45 | |
| ScaLRFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 49.16 | |
| HilDAFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 48.39 | |
| HilDA†Fine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 48.26 | |
| RandomFine-tuning ratio=10%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 47.6 | |
| LiMAFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 45.9 | |
| CleverDFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 44.03 | |
| LiMoEFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 43.95 | |
| SuperFlowFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 42.7 | |
| SealFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 40.64 | |
| SLidRFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 39.6 | |
| ScaLRFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 36.45 | |
| RandomFine-tuning ratio=1%, Backbone=ViT-B/MinkUnet34, Pre-training dataset=nuScenes2026.06 | 23.81 |