ValueError when performing matmul with Tensorflow

Question

I'm a total beginner to TensorFlow, and I'm trying to multiply two matrices together, but I keep getting an exception that says:

ValueError: Shapes TensorShape([Dimension(2)]) and TensorShape([Dimension(None), Dimension(None)]) must have the same rank

Here's minimal example code:

data = np.array([0.1, 0.2])
x = tf.placeholder("float", shape=[2])
T1 = tf.Variable(tf.ones([2,2]))
l1 = tf.matmul(T1, x)
init = tf.initialize_all_variables()

with tf.Session() as sess:
    sess.run(init)
    sess.run(feed_dict={x: data}

Confusingly, the following very similar code works fine:

data = np.array([0.1, 0.2])
x = tf.placeholder("float", shape=[2])
T1 = tf.Variable(tf.ones([2,2]))
init = tf.initialize_all_variables()

with tf.Session() as sess:
    sess.run(init)
    sess.run(T1*x, feed_dict={x: data}

Can anyone point to what the issue is? I must be missing something obvious here..

mrry · Accepted Answer · 2017-11-16T07:11:44.603

20

The tf.matmul() op requires that both of its inputs are matrices (i.e. 2-D tensors)^*, and doesn't perform any automatic conversion. Your T1 variable is a matrix, but your x placeholder is a length-2 vector (i.e. a 1-D tensor), which is the source of the error.

By contrast, the * operator (an alias for tf.multiply()) is a broadcasting element-wise multiplication. It will convert the vector argument to a matrix by following NumPy broadcasting rules.

To make your matrix multiplication work, you can either require that x is a matrix:

data = np.array([[0.1], [0.2]])
x = tf.placeholder(tf.float32, shape=[2, 1])
T1 = tf.Variable(tf.ones([2, 2]))
l1 = tf.matmul(T1, x)
init = tf.initialize_all_variables()

with tf.Session() as sess:
    sess.run(init)
    sess.run(l1, feed_dict={x: data})

...or you could use the tf.expand_dims() op to convert the vector to a matrix:

data = np.array([0.1, 0.2])
x = tf.placeholder(tf.float32, shape=[2])
T1 = tf.Variable(tf.ones([2, 2]))
l1 = tf.matmul(T1, tf.expand_dims(x, 1))
init = tf.initialize_all_variables()

with tf.Session() as sess:
    # ...

^* This was true when I posted the answer at first, but now tf.matmul() also supports batched matrix multiplications. This requires both arguments to have at least 2 dimensions. See the documentation for more details.

edited Nov 16 '17 at 07:11

answered Jan 20 '16 at 19:03

mrry

125,488
26
399
400

Thanks, perfect! That error message was pretty cryptic. – homesalad Jan 20 '16 at 19:55
1

That's a good point! I'll send in a patch to improve it in the next version. – mrry Jan 20 '16 at 21:27
2

Thanks - so is it possible to multiply a vector with a matrix to get a vector (not a matrix)? Or do I have to follow this answer and then reshape ? – Mr_and_Mrs_D Apr 07 '17 at 14:38
Now tf.mul became tf.multiply – aerin Sep 26 '17 at 19:10
@ToussaintLouverture Thanks, I've updated the links. – mrry Sep 27 '17 at 03:58
1

Good suggestion. But you are missing `sess.run(l1, feed_dict={x: data}) }`. Though it is small thing, but for beginners' level it may be helpful . – Shyamkkhadka Nov 16 '17 at 07:00
@Shyamkkhadka Good point: I copied that from the question, and didn't notice that it wasn't valid code. Updated the answer per your suggestion. – mrry Nov 16 '17 at 07:12

ValueError when performing matmul with Tensorflow

1 Answers1

Linked