[FIXED] AttributeError: module 'tensorboard.summary' has no attribute 'pr_curve'

Issue

I am trying to replicate the code I found at this page but I am getting the following error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-21-ac81ef861a8b> in <module>

----> 1 train_and_evaluate(classifier)

<ipython-input-20-6a44bf8c2962> in train_and_evaluate(classifier)
     
12     tf.reset_default_graph()
     
13     # Add a PR summary in addition to the summaries that the classifier writes

---> 14     pr = summary_lib.pr_curve('precision_recall', predictions=predictions, labels=y_test.astype(bool), num_thresholds=21)
     
15     with tf.Session() as sess:
     
16         writer = tf.summary.FileWriter(os.path.join(classifier.model_dir, 'eval'), sess.graph)


AttributeError: module 'tensorboard.summary' has no attribute 'pr_curve'

---------------------------------------------------------------------------

I am currently using:

• Windows 10 21H1
• python 3.6
• Tensorflow 1.15.0
• Tensorflow-Estimator 1.15.1
• Tensorboard 2.5.0

Should I install an older version of tensorboard?

Solution

Tensorlfow 1.15.0 has attribute summary_lib.pr_curve.

Take a look at right configurations

keras-applications-1.0.8
tensorboard-1.15.0
tensorflow-1.15.0 tensorflow-estimator-1.15.1

Working sample code

from tensorboard import summary as summary_lib
import tensorflow as tf

labels = tf.constant([False, True, True, False, True], dtype=tf.bool)
predictions = tf.random_uniform(labels.get_shape(), maxval=1.0)
summary_lib.pr_curve(name='foo',
                     predictions=predictions,
                     labels=labels,
                     num_thresholds=11)
merged_summary = tf.summary.merge_all()

with tf.Session() as sess:
  writer = tf.summary.FileWriter('/tmp/logdir', sess.graph)
  for step in range(43):
    writer.add_summary(sess.run(merged_summary), global_step=step)

Answered By - TFer