mindspore.ops.SeLU

class mindspore.ops.SeLU(*args, **kwargs)[source]

Computes SeLU (scaled exponential Linear Unit) of input tensors element-wise.

The activation function is defined as:

\[E_{i} = scale * \begin{cases} x_{i}, &\text{if } x_{i} \geq 0; \cr \text{alpha} * (\exp(x_i) - 1), &\text{otherwise.} \end{cases}\]

where \(alpha\) and \(scale\) are pre-defined constants(\(alpha=1.67326324\) and \(scale=1.05070098\)).

See more details in Self-Normalizing Neural Networks.

Inputs:

input_x (Tensor) - Tensor of shape \((N, *)\), where \(*\) means, any number of additional dimensions, with float16 or float32 data type.

Outputs:

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

Supported Platforms:

Ascend

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

Examples

>>> input_x = Tensor(np.array([[-1.0, 4.0, -8.0], [2.0, -5.0, 9.0]]), mindspore.float32)
>>> selu = ops.SeLU()
>>> output = selu(input_x)
>>> print(output)
[[-1.1113307 4.202804 -1.7575096]
[ 2.101402 -1.7462534 9.456309 ]]