Factory Functions

simplegrad.core.factory.zeros(shape: tuple[int, ...], dtype: str = 'float32', comp_grad: bool | None = None, label: bool | None = None) -> Tensor

Create a tensor filled with zeros.

Parameters:

Name	Type	Description	Default
shape	tuple[int, ...]	Output shape.	required
dtype	str	Data type string. Defaults to "float32".	'float32'
comp_grad	bool | None	Enable gradient tracking. Defaults to the global flag.	None
label	bool | None	Optional name for visualization.	None

Source code in simplegrad/core/factory.py

def zeros(
    shape: tuple[int, ...],
    dtype: str = "float32",
    comp_grad: bool | None = None,
    label: bool | None = None,
) -> Tensor:
    """Create a tensor filled with zeros.

    Args:
        shape: Output shape.
        dtype: Data type string. Defaults to ``"float32"``.
        comp_grad: Enable gradient tracking. Defaults to the global flag.
        label: Optional name for visualization.
    """
    return Tensor(values=np.zeros(shape), dtype=dtype, comp_grad=comp_grad, label=label)

simplegrad.core.factory.ones(shape: tuple[int, ...], dtype: str = 'float32', comp_grad: bool | None = None, label: bool | None = None) -> Tensor

Create a tensor filled with ones.

Parameters:

Name	Type	Description	Default
shape	tuple[int, ...]	Output shape.	required
dtype	str	Data type string. Defaults to "float32".	'float32'
comp_grad	bool | None	Enable gradient tracking. Defaults to the global flag.	None
label	bool | None	Optional name for visualization.	None

Source code in simplegrad/core/factory.py

def ones(
    shape: tuple[int, ...],
    dtype: str = "float32",
    comp_grad: bool | None = None,
    label: bool | None = None,
) -> Tensor:
    """Create a tensor filled with ones.

    Args:
        shape: Output shape.
        dtype: Data type string. Defaults to ``"float32"``.
        comp_grad: Enable gradient tracking. Defaults to the global flag.
        label: Optional name for visualization.
    """
    return Tensor(values=np.ones(shape), dtype=dtype, comp_grad=comp_grad, label=label)

simplegrad.core.factory.normal(shape: tuple[int, ...], dtype: str = 'float32', comp_grad: bool | None = None, label: bool | None = None, mu: int | float = 0, sigma: int | float = 1) -> Tensor

Create a tensor sampled from a normal distribution N(mu, sigma).

Parameters:

Name	Type	Description	Default
shape	tuple[int, ...]	Output shape.	required
dtype	str	Data type string. Defaults to "float32".	'float32'
comp_grad	bool | None	Enable gradient tracking. Defaults to the global flag.	None
label	bool | None	Optional name for visualization.	None
mu	int | float	Distribution mean.	0
sigma	int | float	Distribution standard deviation.	1

Source code in simplegrad/core/factory.py

def normal(
    shape: tuple[int, ...],
    dtype: "str" = "float32",
    comp_grad: bool | None = None,
    label: bool | None = None,
    mu: int | float = 0,
    sigma: int | float = 1,
) -> Tensor:
    """Create a tensor sampled from a normal distribution N(mu, sigma).

    Args:
        shape: Output shape.
        dtype: Data type string. Defaults to ``"float32"``.
        comp_grad: Enable gradient tracking. Defaults to the global flag.
        label: Optional name for visualization.
        mu: Distribution mean.
        sigma: Distribution standard deviation.
    """
    return Tensor(
        values=np.random.normal(size=shape, loc=mu, scale=sigma),
        dtype=dtype,
        comp_grad=comp_grad,
        label=label,
    )

simplegrad.core.factory.uniform(shape: tuple[int, ...], dtype: str = 'float32', comp_grad: bool | None = None, label: bool | None = None, low: int | float = 0, high: int | float = 1) -> Tensor

Create a tensor sampled from a uniform distribution U(low, high).

Parameters:

Name	Type	Description	Default
shape	tuple[int, ...]	Output shape.	required
dtype	str	Data type string. Defaults to "float32".	'float32'
comp_grad	bool | None	Enable gradient tracking. Defaults to the global flag.	None
label	bool | None	Optional name for visualization.	None
low	int | float	Lower bound of the distribution.	0
high	int | float	Upper bound of the distribution.	1

Source code in simplegrad/core/factory.py

def uniform(
    shape: tuple[int, ...],
    dtype: str = "float32",
    comp_grad: bool | None = None,
    label: bool | None = None,
    low: int | float = 0,
    high: int | float = 1,
) -> Tensor:
    """Create a tensor sampled from a uniform distribution U(low, high).

    Args:
        shape: Output shape.
        dtype: Data type string. Defaults to ``"float32"``.
        comp_grad: Enable gradient tracking. Defaults to the global flag.
        label: Optional name for visualization.
        low: Lower bound of the distribution.
        high: Upper bound of the distribution.
    """
    return Tensor(
        values=np.random.uniform(size=shape, low=low, high=high),
        dtype=dtype,
        comp_grad=comp_grad,
        label=label,
    )

simplegrad.core.factory.full(shape: tuple[int, ...], fill_value: float, dtype: str = 'float32', comp_grad: bool | None = None, label: bool | None = None) -> Tensor

Create a tensor filled with a constant value.

Parameters:

Name	Type	Description	Default
shape	tuple[int, ...]	Output shape.	required
fill_value	float	Value to fill the tensor with.	required
dtype	str	Data type string. Defaults to "float32".	'float32'
comp_grad	bool | None	Enable gradient tracking. Defaults to the global flag.	None
label	bool | None	Optional name for visualization.	None

Source code in simplegrad/core/factory.py

def full(
    shape: tuple[int, ...],
    fill_value: float,
    dtype: str = "float32",
    comp_grad: bool | None = None,
    label: bool | None = None,
) -> Tensor:
    """Create a tensor filled with a constant value.

    Args:
        shape: Output shape.
        fill_value: Value to fill the tensor with.
        dtype: Data type string. Defaults to ``"float32"``.
        comp_grad: Enable gradient tracking. Defaults to the global flag.
        label: Optional name for visualization.
    """
    return Tensor(values=np.full(shape, fill_value), dtype=dtype, comp_grad=comp_grad, label=label)