# tensorplay.sparse.spdiags

Source: https://www.tensorplay.cn/docs/generated/tensorplay.sparse.spdiags.html

# tensorplay.sparse.spdiags

tensorplay.sparse.spdiags(diagonals, offsets, shape, layout=None)[[source]](../_modules/tensorplay/sparse.html#spdiags)

Constructs a sparse tensor from diagonals, mirroring
torch.sparse.spdiags (and thereby scipy.sparse.spdiags).

Parameters:

- diagonals – matrix of shape (len(offsets), L) (or a single vector); row j holds the values of diagonal offsets[j] , read starting at column max(offset, 0) .

- offsets – int64 sequence of distinct diagonal offsets (0 = main, positive = above, negative = below).

- shape – 2-element (M, N) output size.

- layout – output layout tag — tensorplay.sparse_coo (default) or tensorplay.sparse_csr .