Voxelization

torch_lattice.nn.functional.voxelize.spvoxelize(feats, coords, counts)

Return type:

Tensor

Parameters:

• feats (Tensor)

• coords (Tensor)

• counts (Tensor)

torch_lattice.nn.functional.voxelize.voxelize(points, features, *, batch_indices=None, active_rows=None, voxel_size=1.0, origin=0.0, reduction='mean', stride=1)

Quantize point rows into a sparse voxel tensor.

Return type:

SparseTensor

Parameters:

• points (Tensor)

• features (Tensor)

• batch_indices (Tensor | None)

• active_rows (Tensor | int | None)

• reduction (Literal['sum', 'mean'])

torch_lattice.nn.functional.devoxelize.calc_ti_weights(coords, idx_query, scale=1)

Return type:

Tensor

Parameters:

• coords (Tensor)

• idx_query (Tensor)

• scale (float)

torch_lattice.nn.functional.devoxelize.devoxelize(points, voxels, *, batch_indices=None, point_active_rows=None, voxel_size=1.0, origin=0.0, interpolation='nearest')

Sample sparse voxel features at dense point rows.

Return type:

Tensor

Parameters:

• points (Tensor)

• voxels (SparseTensor)

• batch_indices (Tensor | None)

• point_active_rows (Tensor | int | None)

• interpolation (Literal['nearest', 'linear'])

torch_lattice.nn.functional.devoxelize.spdevoxelize(feats, coords, weights)

Return type:

Tensor

Parameters:

• feats (Tensor)

• coords (Tensor)

• weights (Tensor)