Set the number tf.nn.relu () function is a number greater than 0 is kept constant, 0 is less than 0, the function shown in Fig.
ReLU function is one of the neural network activation function commonly used.
ReLU below is an example:
import tensorflow as tf
v = tf.constant([-3, 5, 6, -6,9])
sess = tf.Session()
print('v的原始值为:', end='')
print(sess.run(v))
print('v的ReLu后的值为:', end='')
print(sess.run(tf.nn.relu(v)))
sess.close()
The output is:
v的原始值为:[-3 5 6 -6 9]
v的ReLu后的值为:[0 5 6 0 9]