Sparse relations

torch_lattice.nn.functional.relation.build_pool_output_coords(coords, *, kernel_size, stride=1, padding=0, dilation=1, spatial_range=None)

Generate convolution-style output support entirely on the input device.

Return type:

Tensor

Parameters:

coords (Tensor)

torch_lattice.nn.functional.relation.build_target_out_in_map(input_coords, target_coords, *, kernel_size, stride=1, padding=0, dilation=1)

Build (N_target, kernel_volume) target-to-input row indices.

Return type:

Tensor

Parameters:

• input_coords (Tensor)

• target_coords (Tensor)

torch_lattice.nn.functional.relation.build_target_transposed_out_in_map(input_coords, target_coords, *, kernel_size, stride=1, padding=0, dilation=1)

Build target-to-input rows for transposed convolution geometry.

Return type:

Tensor

Parameters:

• input_coords (Tensor)

• target_coords (Tensor)

torch_lattice.nn.functional.relation.build_transposed_output_coords(coords, *, kernel_size, stride=1, padding=0, dilation=1)

Generate unique support for transposed convolution geometry.

Return type:

Tensor

Parameters:

coords (Tensor)

torch_lattice.nn.functional.relation.gather_scatter_kmap_from_out_in_map(out_in_map, *, input_size)

Convert a target relation to the gather/scatter kernel-map ABI.

Return type:

dict

Parameters:

• out_in_map (Tensor)

• input_size (int)

torch_lattice.nn.functional.conv.kmap.build_kernel_map(_coords, input_node_num, kernel_size=2, stride=2, padding=0, hashmap_keys=None, hashmap_vals=None, spatial_range=None, mode='hashmap', dataflow=Dataflow.ImplicitGEMM, downsample_mode='spconv', training=False, ifsort=False, generative=False, subm=False, split_mask_num=1, split_mask_num_bwd=1, FOD_fusion=True, IGEMM_center_only=False, inference=False)

Return type:

Dict

Parameters:

• _coords (Tensor)

• input_node_num (int)

• kernel_size (int | Tuple[int, ...])

• stride (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• hashmap_keys (Tensor)

• hashmap_vals (Tensor)

• spatial_range (tuple[int, ...] | None)

• training (bool)

• ifsort (bool)

• generative (bool)

• subm (bool)

• split_mask_num (int)

• split_mask_num_bwd (int)

• FOD_fusion (bool)

• IGEMM_center_only (bool)

• inference (bool)

torch_lattice.nn.functional.conv.kmap.spdownsample(_coords, stride=2, kernel_size=2, padding=0, spatial_range=None, downsample_mode='spconv')

Return type:

Tensor

Parameters:

• _coords (Tensor)

• stride (int | Tuple[int, ...])

• kernel_size (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• spatial_range (Tuple[int, ...] | None)

• downsample_mode (str)

torch_lattice.nn.functional.conv.kmap.spupsample_generative(_coords, stride=2, kernel_size=2, padding=0, spatial_range=None)

Return type:

Tensor

Parameters:

• _coords (Tensor)

• stride (int | Tuple[int, ...])

• kernel_size (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• spatial_range (Tuple[int, ...] | None)

torch_lattice.nn.functional.conv.kmap.transpose_kernel_map(kmap, ifsort=False, training=False, split_mask_num=1, split_mask_num_bwd=1)

Return type:

Dict

Parameters:

• kmap (Dict)

• ifsort (bool)

• training (bool)

• split_mask_num (int)

• split_mask_num_bwd (int)

torch_lattice.nn.functional.conv.kmap.build_kmap.build_kernel_map(_coords, input_node_num, kernel_size=2, stride=2, padding=0, hashmap_keys=None, hashmap_vals=None, spatial_range=None, mode='hashmap', dataflow=Dataflow.ImplicitGEMM, downsample_mode='spconv', training=False, ifsort=False, generative=False, subm=False, split_mask_num=1, split_mask_num_bwd=1, FOD_fusion=True, IGEMM_center_only=False, inference=False)

Return type:

Dict

Parameters:

• _coords (Tensor)

• input_node_num (int)

• kernel_size (int | Tuple[int, ...])

• stride (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• hashmap_keys (Tensor)

• hashmap_vals (Tensor)

• spatial_range (tuple[int, ...] | None)

• training (bool)

• ifsort (bool)

• generative (bool)

• subm (bool)

• split_mask_num (int)

• split_mask_num_bwd (int)

• FOD_fusion (bool)

• IGEMM_center_only (bool)

• inference (bool)

torch_lattice.nn.functional.conv.kmap.build_kmap.transpose_kernel_map(kmap, ifsort=False, training=False, split_mask_num=1, split_mask_num_bwd=1)

Return type:

Dict

Parameters:

• kmap (Dict)

• ifsort (bool)

• training (bool)

• split_mask_num (int)

• split_mask_num_bwd (int)

torch_lattice.nn.functional.conv.kmap.downsample.spdownsample(_coords, stride=2, kernel_size=2, padding=0, spatial_range=None, downsample_mode='spconv')

Return type:

Tensor

Parameters:

• _coords (Tensor)

• stride (int | Tuple[int, ...])

• kernel_size (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• spatial_range (Tuple[int, ...] | None)

• downsample_mode (str)

torch_lattice.nn.functional.conv.kmap.upsample.spupsample_generative(_coords, stride=2, kernel_size=2, padding=0, spatial_range=None)

Return type:

Tensor

Parameters:

• _coords (Tensor)

• stride (int | Tuple[int, ...])

• kernel_size (int | Tuple[int, ...])

• padding (int | Tuple[int, ...])

• spatial_range (Tuple[int, ...] | None)