Invalid AUC curve value "%s".

Package:

tensorflow

158813

Exception Class:

ValueError

Raise code

  @staticmethod
  def from_str(key):
    if key in ('pr', 'PR'):
      return AUCCurve.PR
    elif key in ('roc', 'ROC'):
      return AUCCurve.ROC
    else:
      raise ValueError('Invalid AUC curve value "%s".' % key)


class AUCSummationMethod(Enum):
  """Type of AUC summation method.

  https://en.wikipedia.org/wiki/Riemann_sum)
 """

Links to the raise (1)

https://github.com/tensorflow/tensorflow/blob/f26800a1e5b1199cfc1a5aca916edc836b541687/tensorflow/python/keras/utils/metrics_utils.py#L218

Ways to fix

When instantiating AUC class to calculate approximate AUC (Area under the curve) of the ROC or PR curves the valid values for parameter curve are ROC or PR.

Reproducing the error:

pipenv install tensorflow

import tensorflow as tf
m = tf.keras.metrics.AUC(num_thresholds=3,
                         curve='RC' # typo error in RC. it should be ROC
                         )
m.update_state([0, 0, 1, 1], [0, 0.5, 0.3, 0.9])
print(m.result().numpy())

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-38-940549c09add> in <module>()
      1 import tensorflow as tf
----> 2 m = tf.keras.metrics.AUC(num_thresholds=3,curve='RC')
      3 m.update_state([0, 0, 1, 1], [0, 0.5, 0.3, 0.9])
      4 
      5 print(m.result().numpy())

/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/metrics.py in __init__(self, num_thresholds, curve, summation_method, name, dtype, thresholds, multi_label, num_labels, label_weights, from_logits)
   2101       self.curve = curve
   2102     else:
-> 2103       self.curve = metrics_utils.AUCCurve.from_str(curve)
   2104     if isinstance(summation_method, metrics_utils.AUCSummationMethod):
   2105       self.summation_method = summation_method

/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/utils/metrics_utils.py in from_str(key)
    214       return AUCCurve.ROC
    215     else:
--> 216       raise ValueError('Invalid AUC curve value "%s".' % key)
    217 
    218 

ValueError: Invalid AUC curve value "RC".

Fix

The valid value are ROC and PR.

import tensorflow as tf
m = tf.keras.metrics.AUC(num_thresholds=3,
                         curve='ROC'
                         )
m.update_state([0, 0, 1, 1], [0, 0.5, 0.3, 0.9])
print(m.result().numpy())