qubosolver.matrix

qubosolver.Matrix module-attribute

Matrix = Matrixf

2-D float tensor using the globally configured precision (float32 by default).

qubosolver.matrix

Square matrix utilities for QUBO solvers.

A Matrix is a 2-D tensor of shape (n, n) using the globally configured float dtype (float32 by default, float64 when double precision is enabled). This module provides factory functions for creating and converting such matrices on the globally configured torch device.

Typical usage:

Q = matrix.zeros(4)                          # 4x4 zero matrix
Q = matrix.tensor([[0, 1], [1, 0]])          # from nested list
Q = matrix.as_tensor(some_tensor)            # cast existing tensor, no copy when possible

Functions:

• as_tensor –

  Convenience wrapper for torch.as_tensor that converts data to a matrix tensor.

• device –

  Returns the globally configured torch device.

• dtype –

  Returns the globally configured float dtype.

• tensor –

  Creates a matrix tensor from the given data.

• zeros –

  Creates a zero-filled square matrix of shape (n, n).

• zeros_field –

  Creates a dataclass field defaulting to a zero-filled square matrix.

as_tensor

as_tensor(data: Any) ->             qubosolver.Matrix



      module-attribute
   (qubosolver.types.linalg.Matrix)" href="#qubosolver.Matrix">Matrix

Convenience wrapper for torch.as_tensor that converts data to a matrix tensor.

Avoids a copy when possible. If data is already a tensor with the right dtype and on the right device, it is returned as-is, sharing the same underlying memory. A numpy array is also shared rather than copied if it already has the global float dtype and the global device is cpu (numpy arrays only live on CPU, so any other dtype or device forces a copy). Lists, tuples, and other array-like inputs are always copied.

Parameters:

• data (Any (external)) –

  Input data (tensor, numpy array, nested list, etc.).

Returns:

• Matrix –

  A 2-D tensor on the global dtype and device.

Source code in qubosolver/types/matrix.py

def as_tensor(data: Any) -> Matrix:  # noqa: ANN401 (array-like input forwarded to torch.as_tensor)
    """Convenience wrapper for `torch.as_tensor` that converts data to a matrix tensor.

    Avoids a copy when possible. If *data* is already a tensor with the right dtype and on
    the right device, it is returned as-is, sharing the same underlying memory. A numpy
    array is also shared rather than copied if it already has the global float dtype and
    the global device is ``cpu`` (numpy arrays only live on CPU, so any other dtype or
    device forces a copy). Lists, tuples, and other array-like inputs are always copied.

    Args:
        data: Input data (tensor, numpy array, nested list, etc.).

    Returns:
        A 2-D tensor on the global dtype and device.
    """
    return torch.as_tensor(data, dtype=dtype(), device=device())

device

device() -> torch.device

Returns the globally configured torch device.

Source code in qubosolver/types/matrix.py

def device() -> torch.device:
    """Returns the globally configured torch device."""
    return linalg.device()

dtype

dtype() -> torch.dtype

Returns the globally configured float dtype.

Source code in qubosolver/types/matrix.py

def dtype() -> torch.dtype:
    """Returns the globally configured float dtype."""
    return linalg.dtype()

tensor

tensor(data: Any, *, dtype: torch.dtype | None = None, device: torch.device | None = None, **kwargs: Any) -> torch.Tensor

Creates a matrix tensor from the given data.

Parameters:

• data (Any (external)) –

  Input data (nested list or 2-D array-like).

• dtype (torch (external).dtype (external) | None, default: None ) –

  Data type of the tensor.

• device (torch (external).device (external) | None, default: None ) –

  Torch device for the tensor.

• **kwargs (Any (external), default: {} ) –

  Extra keyword arguments forwarded to torch.tensor.

Returns:

• torch (external).Tensor (external) –

  A 2-D tensor.

Source code in qubosolver/types/matrix.py

def tensor(
    data: Any,
    *,
    dtype: torch.dtype | None = None,
    device: torch.device | None = None,
    **kwargs: Any,
) -> torch.Tensor:
    """Creates a matrix tensor from the given data.

    Args:
        data: Input data (nested list or 2-D array-like).
        dtype: Data type of the tensor.
        device: Torch device for the tensor.
        **kwargs: Extra keyword arguments forwarded to `torch.tensor`.

    Returns:
        A 2-D tensor.
    """
    dtype = dtype or _dtype()
    device = device or _device()
    result = torch.tensor(data, dtype=dtype, device=device, **kwargs)
    # `torch.tensor([])` is 1-D: keep an empty matrix 2-D.
    return result.reshape(0, 0) if result.shape == (0,) else result

zeros

zeros(n: int, *, dtype: torch.dtype | None = None, device: torch.device | None = None) -> torch.Tensor

Creates a zero-filled square matrix of shape (n, n).

Parameters:

• n (int (external)) –

  Size of the matrix (number of rows and columns).

• dtype (torch (external).dtype (external) | None, default: None ) –

  Data type of the tensor.

• device (torch (external).device (external) | None, default: None ) –

  Torch device for the tensor.

Returns:

• torch (external).Tensor (external) –

  A 2-D tensor of zeros with shape (n, n).

Source code in qubosolver/types/matrix.py

def zeros(
    n: int, *, dtype: torch.dtype | None = None, device: torch.device | None = None
) -> torch.Tensor:
    """Creates a zero-filled square matrix of shape ``(n, n)``.

    Args:
        n: Size of the matrix (number of rows and columns).
        dtype: Data type of the tensor.
        device: Torch device for the tensor.

    Returns:
        A 2-D tensor of zeros with shape ``(n, n)``.
    """
    dtype = dtype or _dtype()
    device = device or _device()
    return torch.zeros((n, n), dtype=dtype, device=device)

zeros_field

zeros_field(n: int, *, dtype: torch.dtype | None = None, device: torch.device | None = None) ->             qubosolver.Matrix



      module-attribute
   (qubosolver.types.linalg.Matrix)" href="#qubosolver.Matrix">Matrix

Creates a dataclass field defaulting to a zero-filled square matrix.

Parameters:

• n (int (external)) –

  Size of the matrix (number of rows and columns).

• dtype (torch (external).dtype (external) | None, default: None ) –

  Data type of the tensor.

• device (torch (external).device (external) | None, default: None ) –

  Torch device for the tensor.

Returns:

• Matrix –

  A dataclass field (typed as Matrix for the enclosing class) whose

• Matrix –

  default_factory builds a fresh zero tensor per instance.

Source code in qubosolver/types/matrix.py

@no_runtime_typecheck
def zeros_field(
    n: int, *, dtype: torch.dtype | None = None, device: torch.device | None = None
) -> Matrix:
    """Creates a dataclass field defaulting to a zero-filled square matrix.

    Args:
        n: Size of the matrix (number of rows and columns).
        dtype: Data type of the tensor.
        device: Torch device for the tensor.

    Returns:
        A dataclass field (typed as `Matrix` for the enclosing class) whose
        `default_factory` builds a fresh zero tensor per instance.
    """
    return field(default_factory=lambda: zeros(n, dtype=dtype, device=device))