torch.sparse_bsc_tensor

torch.sparse_bsc_tensor(ccol_indices, row_indices, values, size=None, *, dtype=None, device=None, requires_grad=False) → Tensor

Constructs a sparse tensor in BSC (Block Compressed Sparse Column)) with specified 2-dimensional blocks at the given ccol_indices and row_indices. Sparse matrix multiplication operations in BSC format are typically faster than that for sparse tensors in COO format. Make you have a look at the note on the data type of the indices.

Parameters:

• ccol_indices (array_like) – (B+1)-dimensional array of size (*batchsize, ncolblocks + 1). The last element of each batch is the number of non-zeros. This tensor encodes the index in values and row_indices depending on where the given column starts. Each successive number in the tensor subtracted by the number before it denotes the number of elements in a given column.

• row_indices (array_like) – Row block co-ordinates of each block in values. (B+1)-dimensional tensor with the same length as values.

• values (array_list) – Initial blocks for the tensor. Can be a list, tuple, NumPy ndarray, and other types that represents a (1 + 2 + K)-dimensonal tensor where K is the number of dense dimensions.

• size (list, tuple, torch.Size, optional) – Size of the sparse tensor: (*batchsize, nrows * blocksize[0], ncols * blocksize[1], *densesize) If not provided, the size will be inferred as the minimum size big enough to hold all non-zero blocks.

Keyword Arguments:

• dtype (torch.dtype, optional) – the desired data type of returned tensor. Default: if None, infers data type from values.

• device (torch.device, optional) – the desired device of returned tensor. Default: if None, uses the current device for the default tensor type (see torch.set_default_tensor_type()). device will be the CPU for CPU tensor types and the current CUDA device for CUDA tensor types.

• requires_grad (bool, optional) – If autograd should record operations on the returned tensor. Default: False.

Example::

>>> ccol_indices = [0, 1, 2]
>>> row_indices = [0, 1]
>>> values = [[[1, 2], [3, 4]], [[5, 6], [7, 8]]]
>>> torch.sparse_bsc_tensor(torch.tensor(ccol_indices, dtype=torch.int64),
...                         torch.tensor(row_indices, dtype=torch.int64),
...                         torch.tensor(values), dtype=torch.double)
tensor(ccol_indices=tensor([0, 1, 2]),
       row_indices=tensor([0, 1]),
       values=tensor([[[1., 2.],
                       [3., 4.]],
                      [[5., 6.],
                       [7., 8.]]]), size=(2, 2), nnz=2, dtype=torch.float64,
       layout=torch.sparse_bsc)