CODEWITHSUNDEEP
tensorflow
刘米兰
刘米兰

Reputation: 183

tensorflow typed miss when run my train.py

It will report TypeError: The value of a feed cannot be a tf.Tensor object. Acceptable feed values include Python scalars, strings, lists, or numpy ndarrays. when run my tensorflow code Is there some bug below my code? I have convert tensor type by using feed_dict types.Why it still failed?

 with tf.Session() as sess:
                tf.initialize_all_variables().run()
                coord = tf.train.Coordinator()
                threads = tf.train.start_queue_runners(coord=coord)
                print('Initialized!')
                for step in xrange(150000):
                        data,label = read_data(FLAGS.train_file)
                        feed_dict = {train_data_node: data,
                                     train_labels_node: label}
                        _, l, lr, predictions = sess.run([optimizer, loss_value, learning_rate, train_prediction],feed_dict=feed_dict)

Upvotes: 1

Views: 1273

Answers (1)

nessuno
nessuno

Reputation: 27042

The error says: The value of a feed cannot be a tf.Tensor

Your feed is:

feed_dict = {train_data_node: data,
    train_labels_node: label}

Thus, one (or both) between data & label are tf.Tensor object.

You have to extract the value into this object, obtaining an acceptable value for a feed.

To do this, you have to run (or eval) the object before passing it to the feed.

Tl;dr:

feed_dict = {train_data_node: data.eval(),
    train_labels_node: label.eval()}

Upvotes: 1

Related Questions