e3x.nn.activations.bent_identity
- e3x.nn.activations.bent_identity(x)[source]
Bent identity activation function.
Computes the gated linear activation with:
\[\mathrm{gate}(x) = \frac{\sqrt{x^2 + 1}-1}{2x} + 1\]For scalar inputs, this is equivalent to:
\[\mathrm{bent\_identity}(x) = \frac{\sqrt{x^2 + 1}-1}{2} + x\]
- Parameters:
x (
Union[Float[Array, '... 1 (max_degree+1)**2 num_features'], Float[Array, '... 2 (max_degree+1)**2 num_features']]) – Input features to which the nonlinearity is applied.- Return type:
- Returns:
The result of applying the nonlinearity to the input features.