Softplus#
Softplus - 1#
Version
name: Softplus (GitHub)
domain: main
since_version: 1
function: False
support_level: SupportType.COMMON
shape inference: True
This version of the operator has been available since version 1.
Summary
Softplus takes one input data (Tensor<T>) and produces one output data (Tensor<T>) where the softplus function, y = ln(exp(x) + 1), is applied to the tensor elementwise.
Inputs
X (heterogeneous) - T: 1D input tensor
Outputs
Y (heterogeneous) - T: 1D input tensor
Type Constraints
T in ( tensor(double), tensor(float), tensor(float16) ): Constrain input and output types to float tensors.
Examples