qml.math.grad

grad(f, argnums=0)

Compute the gradient in a jax-like manner for any interface.

Parameters:

• f (Callable) – a function with a single 0-D scalar output

• argnums (Sequence[int] | int) – which arguments to differentiate

Returns:

a function with the same signature as f that returns the gradient.

Return type:

Callable

See also

pennylane.math.jacobian()

Note that this function follows the same design as jax. By default, the function will return the gradient of the first argument, whether or not other arguments are trainable.

>>> import jax, torch, tensorflow as tf
>>> def f(x, y):
...     return  x * y
>>> qml.math.grad(f)(qml.numpy.array(2.0), qml.numpy.array(3.0))
tensor(3., requires_grad=True)
>>> qml.math.grad(f)(jax.numpy.array(2.0), jax.numpy.array(3.0))
Array(3., dtype=float32, weak_type=True)
>>> qml.math.grad(f)(torch.tensor(2.0, requires_grad=True), torch.tensor(3.0, requires_grad=True))
tensor(3.)
>>> qml.math.grad(f)(tf.Variable(2.0), tf.Variable(3.0))
<tf.Tensor: shape=(), dtype=float32, numpy=3.0>

argnums can be provided to differentiate multiple arguments.

>>> qml.math.grad(f, argnums=(0,1))(torch.tensor(2.0, requires_grad=True), torch.tensor(3.0, requires_grad=True))
(tensor(3.), tensor(2.))

Note that the selected arguments must be of an appropriately trainable datatype, or an error may occur.

>>> qml.math.grad(f)(torch.tensor(1.0), torch.tensor(2.))
RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn

code/api/pennylane.math.grad

 

Download Python script

 

Download Notebook

 

View on GitHub