eye

iskra.sparse.eye(n: int, dtype: dtype = torch.float32, device: str | device = 'cpu') → SparseTensor[SOURCE]

Constructs a sparse COO identity matrix.

Parameters:

• n (int) – Size (number of rows or columns) of the matrix.

• dtype (torch.dtype) – Matrix data type. Defaults to torch.float32.

• device (str | torch.device, optional) – Device for the matrix. Defaults to “cpu”.

Returns:

​ (SparseTensor[DType, [n, n]]) – Sparse n-by-n identity matrix in COO format.