tfa.activations.mish

Note: tensorflow-addons is deprecated, use tf.keras.activations.mish instead.

Mish: A Self Regularized Non-Monotonic Neural Activation Function.

tfa.activations.mish(
    x: tfa.types.TensorLike
) -> tf.Tensor

Computes mish activation:

$$\mathrm{mish}(x) = x \cdot \tanh(\mathrm{softplus}(x)).$$

See Mish: A Self Regularized Non-Monotonic Neural Activation Function.

Usage:

x = tf.constant([1.0, 0.0, 1.0])
tfa.activations.mish(x)
<tf.Tensor: shape=(3,), dtype=float32, numpy=array([0.865098..., 0.       , 0.865098...], dtype=float32)>

Args
x	A Tensor. Must be one of the following types: bfloat16, float16, float32, float64.

Returns
A Tensor. Has the same type as x.