tensorflow/models · error · ValueError

Cache is not supported when training is True.

Error message

Cache is not supported when training is True.

What it means

Error "Cache is not supported when training is True." thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/layers/kernel_attention.py:769

    Args:
      query: Query `Tensor` of shape `[B, T, dim]`.
      value: Value `Tensor` of shape `[B, S, dim]`.
      key: Optional key `Tensor` of shape `[B, S, dim]`. If not given, will use
        `value` for both `key` and `value`, which is the most common case.
      attention_mask: a boolean mask of shape `[B, S]`, that prevents attenting
        to masked positions. Note that the mask is only appied to the keys. User
        may want to mask the output if query contains pads.
      cache: Cache to accumulate history in memory. Used at inferecne time
        (streaming, decoding) for  causal attention.
      training: Python boolean indicating whether the layer should behave in
        training mode (adding dropout) or in inference mode (doing nothing).

    Returns:
      Multi-headed outputs of attention computation.
    """
    if cache is not None:
      if training:
        raise ValueError(
            "Cache is not supported when training is True.")
      if not self.use_causal_windowed:
        raise ValueError(
            "Cache is not supported for non use_causal_windowed case.")
      if self._begin_kernel:
        raise ValueError(
            "Cache is not supported when begin_kernel is set since the bahvior "
            "is too complicated.")
      if self._feature_transform in _NON_CAUSAL_SUPPORT_TRANSFORM_MAP:
        raise ValueError("Cache is not supported for feature_transform %s" %
                         (self._feature_transform))

    if not self._built_from_signature:
      self._build_from_signature(query=query, value=value, key=key)
    if key is None:
      key = value

    #   N = `num_attention_heads`

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/layers/kernel_attention.py:769 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/b137b230cbf7609f. Report an issue: GitHub.