Python v0.3.0

LeakyReLU

class dragon.vm.tensorflow.keras.layers.LeakyReLU(
  alpha=0.3,
  **kwargs
)[source]

Layer to apply the leaky rectified linear unit. [Clevert et.al, 2015].

The LeakyReLU function is defined as:

\[\text{LeakyReLU}(x) = \begin{cases} x, & \text{ if } x \geq 0 \\ \alpha * x, & \text{ otherwise } \end{cases} \]

Examples:

x = tf.constant([-1, 0, 1], 'float32')
print(tf.keras.layers.LeakyReLU(inplace=False)(x))

__init__

LeakyReLU.__init__(
  alpha=0.3,
  **kwargs
)[source]

Create a LeakyReLU layer.

Parameters:
  • alpha (float, optional, default=0.3) – The value to \(\alpha\).