tensorflow how to get prediction

Question

I want to compute predictions, using the following code:

import tensorflow as tf

x = tf.placeholder("float", [None, n_input])
y = tf.placeholder("float", [None, n_classes])

pred = multilayer_perceptron(x, weights, biases)


cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(pred, y))
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cost)

# Initializing the variables

##trn.txt start

##tst.txt end
with tf.Session() as sess:
    sess.run(init)

    # Training cycle
    for epoch in range(training_epochs):
        avg_cost = 0.
        total_batch = int(num_lines_trn/batch_size)
        # Loop over all batches
        for i in range(total_batch):
            batch_x, batch_y = bat_x[i*batch_size:(i+1)*batch_size],bat_y[i*batch_size:(i+1)*batch_size]#mnist.train.next_batch(batch_size)
            # Run optimization op (backprop) and cost op (to get loss value)
            _, c = sess.run([optimizer, cost], feed_dict={x: batch_x,
                                                          y: batch_y})
            # Compute average loss
            avg_cost += c / total_batch

    correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))

    accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

    print(sess.run(accuracy, feed_dict={x: tst_x, y: tst_y}))
    print(sess.run(accuracy, feed_dict={x: tst_x}))

The line

print(sess.run(accuracy, feed_dict={x: tst_x, y: tst_y}))

returns 0.80353 which is the accuracy for the batch.

However I want to get prediction result. so I added:

print(sess.run(accuracy, feed_dict={x: tst_x}))

But this line returns an error:

You must feed a value for placeholder tensor 'Placeholder_7' with dtype float

How can I solve this problem?

Answer 1

If you want to get the predictions of your model, you should do:

sess.run(pred, feed_dict={x: tst_x})

You have an error because you try to run sess.run(accuracy, feed_dict={x: tst_x}), but to compute the accuracy on a given batch, you need the true labels contained in placeholder y, so you get the following error:

You must feed a value for placeholder tensor 'Name of placeholder y'