tensorflow/models · error · NotImplementedError

Fused projection is not supported.

Error message

Fused projection is not supported.

What it means

Error "Fused projection is not supported." thrown in tensorflow/models.

Source

Thrown at official/projects/edgetpu/vision/modeling/mobilenet_edgetpu_v2_model_blocks.py:858

    activation = tf_utils.get_activation(config.activation)
    drop_connect_rate = config.drop_connect_rate
    data_format = tf_keras.backend.image_data_format()
    use_depthwise = block.conv_type == 'depthwise'
    use_groupconv = block.conv_type == 'group'
    prefix = prefix or ''
    self.name = prefix
    conv_kernel_initializer = (
        config.conv_kernel_initializer if config.conv_kernel_initializer
        is not None else CONV_KERNEL_INITIALIZER)

    filters = block.input_filters * block.expand_ratio

    self.expand_block: List[tf_keras.layers.Layer] = []
    self.squeeze_excitation: List[tf_keras.layers.Layer] = []
    self.project_block: List[tf_keras.layers.Layer] = []

    if block.fused_project:
      raise NotImplementedError('Fused projection is not supported.')

    if block.fused_expand and block.expand_ratio != 1:
      # If we use fused mbconv, fuse expansion with the main kernel.
      # If conv_type is depthwise we still fuse it to a full conv.
      if use_groupconv:
        self.expand_block.append(groupconv2d_block(
            filters,
            config,
            kernel_size=block.kernel_size,
            strides=block.strides,
            group_size=block.group_size,
            activation=activation,
            name=prefix + 'fused'))
      else:
        self.expand_block.extend(
            conv2d_block_as_layers(
                conv_filters=filters,
                config=config,

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/edgetpu/vision/modeling/mobilenet_edgetpu_v2_model_blocks.py:858 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/0f8c33329d81d8aa. Report an issue: GitHub.