layers

Reusable neural network blocks shared across the model architectures.

Modules:

• act –

  Activation factory wrapper around torch_geometric.nn.resolver.

• affine –

  Per-channel affine transformation as a function and a learnable module.

• anchors –

  Anchor-based dense detection heads for the voxel detectors (PointPillars, SECOND).

• bev_backbone –

  SSD-style 2D BEV backbones shared by the voxel detectors (PointPillars, SECOND, Voxel Mamba).

• conv2d_blocks –

  2D convolution block with optional normalization and activation.

• conv3d_blocks –

  3D convolution block with optional normalization and activation.

• dropouts –

  Dropout layers: stochastic depth (DropPath) and the create_dropout factory.

• fps –

  Farthest point sampling (FPS) as a module returning sampled indices.

• geometric_affine –

  PointMLP grouping convolution with geometric affine neighborhood normalization.

• grid_pool –

  Grid-based point cloud pooling (downsampling).

• linear_blocks –

  Linear block with optional normalization and activation.

• norms –

  Normalization factory wrapper with spatial dimensionality support.

• octree_attention –

  OctFormer octree window attention with relative position encoding.

• octree_blocks –

  Octree convolution and deconvolution blocks with normalization and activation.

• pdnorm –

  Prompt-driven normalization (PDNorm) routing each batch through a per-condition norm.

• point_patch_embed –

  Mini-PointNet patch embedding turning local point groups into tokens.

• pointconv –

  PointConv message-passing convolutions, with and without density reweighting.

• pointconv_sa –

  PointConv set abstraction modules for hierarchical and global feature extraction.

• pointnet2_blocks –

  PointNet++ set abstraction and feature propagation blocks.

• pointnext_blocks –

  PointNeXt convolution layer introduced in the

• pools –

  Per-segment pooling modules over packed batches and the create_pool / create_adaptive_pool factories.

• pvcnn_blocks –

  PVCNN building blocks: voxelization, squeeze-and-excitation, and point-voxel convolution.

• rope –

  3D rotary position embedding for point-cloud attention.

• serialized_attention –

  Serialized attention variants used by Point Transformer V3 and descendants.

• serialized_pool –

  Pooling and upsampling driven by point cloud serialization codes.

• spconv_blocks –

  Sparse convolution blocks: submanifold, strided, and residual variants.

• tnet –

  T-Net alignment modules predicting affine transforms for point and feature spaces.

• transformer –

  Standard pre-norm transformer building blocks for point-cloud backbones.

• vfe –

  Dynamic voxel feature encoder shared by the voxel detectors (Voxel Mamba, LION).

• view –

  Tensor reshaping as a module wrapping Tensor.view.

• xconv –

  PointCNN XConv layer: convolution on X-transformed point neighborhoods.