Selu activation for keras and tensorflow backend, works better than Relu

Selu_keras.py

import keras.backend as K
from keras.engine import Layer
import tensorflow as tf


class SelU(Layer):

    def __init__(self, alpha=1.6732632423543772848170429916717, scale=1.0507009873554804934193349852946, **kwargs):
        super(SelU, self).__init__(**kwargs)
        self.alpha = K.cast_to_floatx(alpha)
        self.scale = K.cast_to_floatx(scale)

    def call(self, inputs, **kwargs):
        return self.scale * tf.where(inputs > 0.0, inputs, self.alpha * tf.exp(inputs) - self.alpha)

    def get_config(self):
        config = {"alpha": float(self.alpha), "scale": float(self.scale)}
        base_config = super(SelU, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))