FIND: Frequency Invariance Disentanglement for Test-Time Adaptation in LiDAR 3D Detection
Paper β’ 2611.09170 β’ Published
Source-pretrained SECOND-IoU checkpoints for FIND: Frequency Invariance Disentanglement for Test-Time Adaptation in LiDAR 3D Detection (NeurIPS 2026).
Paper: arXiv:2611.09170
| Checkpoint | Source dataset | Adaptation task |
|---|---|---|
| n2k.pth | nuScenes | nuScenes β KITTI |
| w2k.pth | Waymo | Waymo β KITTI |
| k2c.pth | KITTI | KITTI β KITTI-C |
Place the downloaded checkpoints in the checkpoints/ directory of the FIND codebase.
Follow the GitHub README for installation, data preparation, source spectral statistics, and online adaptation.