def call()

in tensorflow_recommenders/tasks/ranking.py [0:0]

• 26 lines of code
• 8 McCabe index (conditional complexity)

  def call(self,
           labels: tf.Tensor,
           predictions: tf.Tensor,
           sample_weight: Optional[tf.Tensor] = None,
           training: bool = False,
           compute_metrics: bool = True) -> tf.Tensor:
    """Computes the task loss and metrics.

    Args:
      labels: Tensor of labels.
      predictions: Tensor of predictions.
      sample_weight: Tensor of sample weights.
      training: Indicator whether training or test loss is being computed.
      compute_metrics: Whether to compute metrics. Set this to False
        during training for faster training.

    Returns:
      loss: Tensor of loss values.
    """

    loss = self._loss(
        y_true=labels, y_pred=predictions, sample_weight=sample_weight)

    if not compute_metrics:
      return loss

    update_ops = []

    for metric in self._ranking_metrics:
      update_ops.append(metric.update_state(
          y_true=labels, y_pred=predictions, sample_weight=sample_weight))

    for metric in self._prediction_metrics:
      update_ops.append(
          metric.update_state(predictions, sample_weight=sample_weight))

    for metric in self._label_metrics:
      update_ops.append(
          metric.update_state(labels, sample_weight=sample_weight))

    for metric in self._loss_metrics:
      update_ops.append(
          metric.update_state(loss, sample_weight=sample_weight))

    # Custom metrics may not return update ops, unlike built-in
    # Keras metrics.
    update_ops = [x for x in update_ops if x is not None]

    with tf.control_dependencies(update_ops):
      return tf.identity(loss)