tf.keras.metrics.get

TensorFlow 1 version View source on GitHub

Retrieves a Keras metric as a function/Metric class instance.

The identifier may be the string name of a metric function or class.

metric = tf.keras.metrics.get("categorical_crossentropy")
type(metric)
<class 'function'>
metric = tf.keras.metrics.get("CategoricalCrossentropy")
type(metric)
<class '...tensorflow.python.keras.metrics.CategoricalCrossentropy'>

You can also specify config of the metric to this function by passing dict containing class_name and config as an identifier. Also note that the class_name must map to a Metric class

identifier = {"class_name": "CategoricalCrossentropy",
              "config": {"from_logits": True} }
metric = tf.keras.metrics.get(identifier)
type(metric)
<class '...tensorflow.python.keras.metrics.CategoricalCrossentropy'>

identifier A metric identifier. One of None or string name of a metric function/class or metric configuration dictionary or a metric function or a metric class instance

A Keras metric as a function/ Metric class instance.

ValueError If identifier cannot be interpreted.