Fix tf.matmul() ValueError: Shape must be rank 2 but is rank 3 for ‘MatMul’ – TensorFlow Tutorial

By | October 12, 2020

In this tutorial, we will disucss and fix the tf.matmul() ValueError: Shape must be rank 2 but is rank 3 for ‘MatMul’ (op: ‘MatMul’) in tensorflow.

Look at code below:

import tensorflow as tf
import numpy as np

t3 = tf.Variable(np.array([[200, 4, 5], [20, 5, 70],[2, 3, 5], [5, 5, 7]]), dtype = tf.float32)

w = tf.Variable(tf.random_uniform([3,3], -0.01, 0.01))
t3 = tf.reshape(t3, [2,2,3])
wx = tf.matmul(t3, w)

Here the shape of \(t3\) is [2, 2, 3], the rank is 3

The shape of \(w\) is [3, 3], the rank is 2.

How about \(wx\)?

We will test this code in different tensorflow.

TensorFlow = 1.10.0

init = tf.global_variables_initializer() 
init_local = tf.local_variables_initializer()

with tf.Session() as sess:
    sess.run([init, init_local])
    print(sess.run([wx]))

Run this code, you will get error.

Fix tf.matmul() ValueError - Shape must be rank 2 but is rank 3 for 'MatMul' - TensorFlow Tutorial

However, we run this code in tensorflow 1.14.0

TensorFlow = 1.14.0

Run code above, we will get the result:

[[[ 0.371 -1.872  0.757]
  [-0.379 -0.345 -0.372]]

 [[-0.046 -0.024 -0.043]
  [-0.066 -0.052 -0.057]]]

There is no error in tensorflow 1.14.0, which means if you plan to avoid this error, you can update your tensorflow version.

Check your current tensorflow version:

Print TensorFlow Version for Beginners – TensorFlow Tutorial

Update or Install specific tensorflow version:

Conda Install Specific TensorFLow Version: A Completed Guide – TensorFlow Tutorial

Leave a Reply

Your email address will not be published. Required fields are marked *