NN utilities

torch_lattice.nn.utils.apply.fapply(input, fn, *args, **kwargs)

Return type:

SparseTensor

Parameters:

• input (SparseTensor)

• fn (Callable[[...], Tensor])

torch_lattice.nn.utils.kernel.get_kernel_offsets(size, stride=1, dilation=1, device='cpu')

Return type:

Tensor

Parameters:

• size (int | Tuple[int, ...])

• stride (int | Tuple[int, ...])

• dilation (int | Tuple[int, ...])