Missed TensorFlow World? Check out the recap. Learn more

tf.keras.losses.poisson

View source on GitHub

Computes the Poisson loss between y_true and y_pred.

Aliases:

  • tf.compat.v1.keras.losses.poisson
  • tf.compat.v1.keras.metrics.poisson
  • tf.compat.v2.keras.losses.poisson
  • tf.compat.v2.keras.metrics.poisson
  • tf.compat.v2.losses.poisson
  • tf.compat.v2.metrics.poisson
  • tf.keras.metrics.poisson
  • tf.losses.poisson
  • tf.metrics.poisson
tf.keras.losses.poisson(
    y_true,
    y_pred
)

The Poisson loss is the mean of the elements of the Tensor y_pred - y_true * log(y_pred).

Usage:

loss = tf.keras.losses.poisson([1.4, 9.3, 2.2], [4.3, 8.2, 12.2])
print('Loss: ', loss.numpy())  # Loss: -0.8045559

Args:

  • y_true: Tensor of true targets.
  • y_pred: Tensor of predicted targets.

Returns:

A Tensor with the mean Poisson loss.

Raises:

  • InvalidArgumentError: If y_true and y_pred have incompatible shapes.