CODEWITHSUNDEEP
tensorflow2.0
RAJARSHI BHOWMIK
RAJARSHI BHOWMIK

Reputation: 53

What is the replacement for tf.contrib.layers.layer_norm() in TensorFlow 2.0?

It seems that TF 2.0 does not have tf.contrib. Then, what is the replacement for tf.contrib.layers.layer_norm()? I didn't find it in tensorflow_addons too.

Upvotes: 5

Views: 4317

Answers (2)

Prashant
Prashant

Reputation: 1

As everyone explained the new version of tensorflow above 2.0 does not support contrib. The simplest method to resolve this problem is go the file where it is showing error and change tf.contrib.layers to tf.keras.layers. It worked for me.

Upvotes: 0

Shanqing Cai
Shanqing Cai

Reputation: 3876

The replacement in the core TensorFlow 2.0 API is tf.keras.lyaers.LayerNormalizaton().

Please see the documentation at: https://www.tensorflow.org/api_docs/python/tf/keras/layers/LayerNormalization?version=stable.

Below is a simple usage example:

import numpy as np
import tensorflow as tf

model = tf.keras.Sequential()
model.add(tf.keras.layers.Dense(10, input_shape=[20], activation="relu"))
# Here is the LayerNormalization layer.
model.add(tf.keras.layers.LayerNormalization())
model.add(tf.keras.layers.Dense(1, activation="sigmoid"))
model.summary()

print(model.predict(np.ones([1, 20])))

Upvotes: 2

Related Questions