tensorflow/models · error · ValueError

The bottleneck size {intra_bottleneck_size} is not a multipl

Error message

The bottleneck size {intra_bottleneck_size} is not a multiple of the number of attention heads {num_attention_heads}.

What it means

Error "The bottleneck size {intra_bottleneck_size} is not a multiple of the number of attention heads {num_attention_heads}." thrown in tensorflow/models.

Source

Thrown at official/projects/qat/nlp/modeling/layers/mobile_bert_layers.py:280

    Raises:
      ValueError: A Tensor shape or parameter is invalid.
    """
    super().__init__(**kwargs)
    self.hidden_size = hidden_size
    self.num_attention_heads = num_attention_heads
    self.intermediate_size = intermediate_size
    self.intermediate_act_fn = intermediate_act_fn
    self.hidden_dropout_prob = hidden_dropout_prob
    self.attention_probs_dropout_prob = attention_probs_dropout_prob
    self.intra_bottleneck_size = intra_bottleneck_size
    self.use_bottleneck_attention = use_bottleneck_attention
    self.key_query_shared_bottleneck = key_query_shared_bottleneck
    self.num_feedforward_networks = num_feedforward_networks
    self.normalization_type = normalization_type
    self.initializer = tf_keras.initializers.get(initializer)

    if intra_bottleneck_size % num_attention_heads != 0:
      raise ValueError(
          (f'The bottleneck size {intra_bottleneck_size} is not a multiple '
           f'of the number of attention heads {num_attention_heads}.'))
    attention_head_size = int(intra_bottleneck_size / num_attention_heads)

    self.block_layers = {}
    # add input bottleneck
    dense_layer_2d = _quantized_einsum_dense(
        'abc,cd->abd',
        output_shape=[None, self.intra_bottleneck_size],
        bias_axes='d',
        kernel_initializer=initializer,
        name='bottleneck_input/dense')
    layer_norm = _output_quantize(
        _get_norm_layer(self.normalization_type,
                        name='bottleneck_input/norm'))
    self.block_layers['bottleneck_input'] = [dense_layer_2d,
                                             layer_norm]

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/qat/nlp/modeling/layers/mobile_bert_layers.py:280 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/70cb5210b1d78461. Report an issue: GitHub.