mindspore.nn.FastGelu

class mindspore.nn.FastGelu[source]

Fast Gaussian error linear unit activation function.

Applies FastGelu function to each element of the input. The input is a Tensor with any valid shape.

FastGelu is defined as:

\[FastGelu(x_i) = \frac {x_i} {1 + \exp(-1.702 * \left| x_i \right|)} * \exp(0.851 * (x_i - \left| x_i \right|))\]

where \(x_i\) is the element of the input.

Inputs:

x (Tensor) - The input of FastGelu with data type of float16 or float32. The shape is \((N,*)\) where \(*\) means, any number of additional dimensions.

Outputs:

Tensor, with the same type and shape as the x.

Raises: TypeError – If dtype of x is neither float16 nor float32.

Supported Platforms:: Ascend

Examples

>>> x = Tensor(np.array([[-1.0, 4.0, -8.0], [2.0, -5.0, 9.0]]), mindspore.float32)
>>> fast_gelu = nn.FastGelu()
>>> output = fast_gelu(x)
>>> print(output)
[[-1.5418735e-01  3.9921875e+00 -9.7473649e-06]
 [ 1.9375000e+00 -1.0052517e-03  8.9824219e+00]]