tensorflow: How to feed numpy.ndarray?

Question

I decoded a JPEG image and have it in the shape n_samples x n_features as a two-dimensional numpy.ndarray. I feed this to tensorflow as following:

sess.run(train_step, feed_dict={X : train_set.Features, y_ : train_set.labels})

This returns a TypeError: TypeError: unhashable type: 'numpy.ndarray'.

I think it is a simple issue, but I cannot find good advise on that. Closest I found was this post on stack overflow but as far as I understand, that's what I do.

Can you show us the X, train_set.Features, y_, train_set.labels (and their shapes)? Some of them probably have the wrong shape. — Sung Kim
– Sung Kim, Commented Apr 29, 2016 at 22:46
Probably your X or y_ is a numpy array, but it should be tf.Tensor or str containing name of endpoint — Yaroslav Bulatov
– Yaroslav Bulatov, Commented Apr 29, 2016 at 23:23

xuesen Zhang · Accepted Answer · 2016-04-30 12:58:16Z

1

I guess your X and train_set.Features maybe have different shape. for examples,

# cifar10 datasets 
x = tf.placeholder(tf.float32,shape = (None,32,32,3))
y = tf.placeholder(tf.float32,shape = (None,10))
print x_batch.shape   # (batch_size,32,32,3)
print y_batch.shape   # (batch_size,10)
# and feed_dict should be
feed_dict = {x:x_batch,y:y_batch}

answered Apr 30, 2016 at 12:58

xuesen Zhang

411 silver badge5 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

tensorflow: How to feed numpy.ndarray?

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related