mygrad.nnet.initializers.normal#

mygrad.nnet.initializers.normal(*shape, mean=0, std=1, dtype=<class 'numpy.float32'>, constant=None)[source]#

Initialize a mygrad.Tensor by drawing from a normal (Gaussian) distribution.

Parameters:

shapeSequence[int]

The output shape.

meanReal, optional (default=0)

The mean of the distribution from which to draw.

stdReal, optional (default=1)

The standard deviation of the distribution from which to draw.

dtypedata-type, optional (default=float32)

The data type of the output tensor; must be a floating-point type.

constantbool, optional (default=False)

If True, the returned tensor is a constant (it: does not back-propagate a gradient).

Returns:

mygrad.Tensor, shape=``shape``: A Tensor, with values drawn from Ɲ(μ, σ²), where μ=``mean`` and σ=``std``.

Examples

>>> from mygrad.nnet.initializers import normal
>>> normal(1, 2, 3)
Tensor([[[-0.06481607, -0.550582  ,  0.04689528],
         [ 0.82973075,  2.83742   ,  1.0964519 ]]], dtype=float32)

>>> normal(2, 2, dtype="float16", constant=True)
Tensor([[-1.335 ,  0.9297],
        [ 1.746 , -0.1222]], dtype=float16)

>>> normal(5, dtype="float64")
Tensor([-0.03875407,  0.65368466, -0.72636993,  1.57404148, -1.17444345])