tensorflow/models · error · ValueError

Invalid activation function string is passed: %s

Error message

Invalid activation function string is passed: %s

What it means

Error "Invalid activation function string is passed: %s" thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/models/t5.py:427

  def __init__(self,
               d_model: int,
               d_ff: int,
               activations: Sequence[str],
               use_bias: bool = False,
               dropout_rate: Optional[float] = 0.0,
               layer_norm_epsilon: Optional[float] = 1e-6,
               weight_initializer: Optional[Initializer] = None,
               bias_initializer: Optional[Initializer] = None,
               **kwargs):
    super().__init__(**kwargs)
    self.use_bias = use_bias
    with self.name_scope:
      self.wi = []
      self.activations = activations
      for idx, act_fn in enumerate(activations):
        if (act_fn is not None and act_fn != "linear" and
            act_fn not in self.activation_map):
          raise ValueError("Invalid activation function string is passed: %s" %
                           act_fn)
        dense_name = "wi" if len(activations) == 1 else f"wi_{idx}"
        self.wi.append(
            Linear(
                d_model,
                d_ff,
                use_bias=self.use_bias,
                w_init=weight_initializer,
                b_init=bias_initializer,
                dtype=self.dtype,
                name=dense_name))

      self.wo = Linear(
          d_ff,
          d_model,
          use_bias=self.use_bias,
          w_init=weight_initializer,
          b_init=bias_initializer,

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/models/t5.py:427 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/d60fb36e029195a1. Report an issue: GitHub.