tensorflow/models · error · ValueError

The treatment_indicator feature (specified as '{self._treatm

Error message

The treatment_indicator feature (specified as '{self._treatment_indicator_feature_name}') must be part of the inputs during training and evaluation, but got input features {set(inputs.keys())} instead.

What it means

Error "The treatment_indicator feature (specified as '{self._treatment_indicator_feature_name}') must be part of the inputs during training and evaluation, but got input features {set(inputs.keys())} instead." thrown in tensorflow/models.

Source

Thrown at official/recommendation/uplift/models/two_tower_uplift_model.py:80

    self._output_head = two_tower_output_head.TwoTowerOutputHead(
        treatment_indicator_feature_name=treatment_indicator_feature_name,
        uplift_network=uplift_network,
        inverse_link_fn=inverse_link_fn,
    )

  def call(
      self,
      inputs: types.DictOfTensors,
      training: bool | None = None,
      mask: tf.Tensor | None = None,
  ) -> types.TwoTowerPredictionOutputs | types.TwoTowerTrainingOutputs:
    return self._output_head(inputs=inputs, training=training, mask=mask)

  def _assert_treatment_indicator_in_data(self, data):
    inputs, _, _ = tf_keras.utils.unpack_x_y_sample_weight(data)

    if self._treatment_indicator_feature_name not in inputs:
      raise ValueError(
          "The treatment_indicator feature (specified as"
          f" '{self._treatment_indicator_feature_name}') must be part of the"
          " inputs during training and evaluation, but got input features"
          f" {set(inputs.keys())} instead."
      )

  def train_step(self, data) -> types.TwoTowerTrainingOutputs:
    self._assert_treatment_indicator_in_data(data)
    return super().train_step(data)

  def test_step(self, data) -> types.TwoTowerTrainingOutputs:
    self._assert_treatment_indicator_in_data(data)
    return super().test_step(data)

  def predict_step(self, data) -> dict[str, tf.Tensor]:
    outputs = super().predict_step(data)

    return {

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/recommendation/uplift/models/two_tower_uplift_model.py:80 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24). Data as JSON: /api/errors/f6b36462a6996df6. Report an issue: GitHub.