| --- |
| license: mit |
| library_name: pytorch |
| tags: |
| - lidar |
| - point-cloud |
| - point-cloud-completion |
| - scene-completion |
| - 3d |
| - semantickitti |
| - autonomous-driving |
| - pytorch_model_hub_mixin |
| - model_hub_mixin |
| --- |
| |
| # RapidLiDAR |
|
|
| Single-pass LiDAR scene completion trained on SemanticKITTI. |
|
|
| Code, installation, and full documentation: [https://github.com/AzharSindhi/RapidLiDAR](https://github.com/AzharSindhi/RapidLiDAR) |
|
|
| Arxiv: [http://arxiv.org/abs/2608.16490](http://arxiv.org/abs/2608.16490) |
|
|
| ## Usage |
|
|
| ```python |
| from rapidlidar.models.hub import RapidLiDARHubModel |
| |
| model = RapidLiDARHubModel.from_pretrained("Azhar88/RapidLiDAR-coarse") |
| model.eval() |
| |
| # x_partial: (B, N, 3) partial point cloud |
| output = model(x_partial, up_factor=10) |
| completed_points = output.points # (B, N * up_factor, 3) |
| ``` |
