array

array#

Signature#

nabla.array(data: 'list | np.ndarray | float | int', dtype: 'DType', device: 'Device', batch_dims: 'Shape', traced: 'bool') -> 'Array'

Description#

Creates an array from a Python list, NumPy array, or scalar.

This function is the primary way to create a Nabla array from existing data. It converts the input data into a Nabla array on the specified device and with the given data type.

Parameters#

  • data (list | np.ndarray | float | int): The input data to convert to an array.

  • dtype (DType, optional): The desired data type for the array. Defaults to DType.float32.

  • device (Device, optional): The computational device where the array will be stored. Defaults to the CPU.

  • batch_dims (Shape, optional): Specifies leading dimensions to be treated as batch dimensions. Defaults to an empty tuple.

  • traced (bool, optional): Whether the operation should be traced in the graph. Defaults to False.

Returns#

  • Array: A new Nabla array containing the provided data.

Examples#

>>> import nabla as nb
>>> import numpy as np
>>> # Create from a Python list
>>> nb.array([1, 2, 3])
Array([1, 2, 3], dtype=int32)

>>> # Create from a NumPy array
>>> np_arr = np.array([[4.0, 5.0], [6.0, 7.0]])
>>> nb.array(np_arr)
Array([[4., 5.],
       [6., 7.]], dtype=float32)

>>> # Create a scalar array
>>> nb.array(100, dtype=nb.DType.int64)
Array(100, dtype=int64)
Contents