tensorflow/models · error · ValueError

The last dimension of the inputs to `TNExpandCondense` shoul

Error message

The last dimension of the inputs to `TNExpandCondense` should be defined. Found `None`.

What it means

Error "The last dimension of the inputs to `TNExpandCondense` should be defined. Found `None`." thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/layers/tn_expand_condense.py:85

      kwargs['input_shape'] = (kwargs.pop('input_dim'),)

    super().__init__(**kwargs)

    assert proj_multiplier in [
        2, 4, 6, 8, 10, 12
    ], 'proj_multiplier needs to be one of [2, 4, 6, 8, 10, 12]'
    self.proj_multiplier = proj_multiplier

    self.use_bias = use_bias
    self.activation = activations.get(activation)
    self.kernel_initializer = initializers.get(kernel_initializer)
    self.bias_initializer = initializers.get(bias_initializer)

  def build(self, input_shape: List[int]) -> None:
    # Disable the attribute-defined-outside-init violations in this function
    # pylint: disable=attribute-defined-outside-init
    if input_shape[-1] is None:
      raise ValueError(
          'The last dimension of the inputs to `TNExpandCondense` '
          'should be defined. Found `None`.')

    super().build(input_shape)

    self.proj_size = self.proj_multiplier * input_shape[-1]

    assert (self.proj_size // input_shape[-1]) * input_shape[
        -1] == self.proj_size, (f'{self.proj_size} / {input_shape[-1]} must be '
                                f'round')
    assert (input_shape[-1] // 128
           ) * 128 == input_shape[-1], f'{input_shape[-1]} / 128 must be round'

    self.w1 = self.add_weight(
        name='w1',
        shape=(input_shape[-1], input_shape[-1]),
        trainable=True,
        initializer=tf_utils.clone_initializer(self.kernel_initializer))

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/layers/tn_expand_condense.py:85 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/dc939b9148cad223. Report an issue: GitHub.