tensorflow/models · error · ValueError

unknown combine_fn `{}`.

Error message

unknown combine_fn `{}`.

What it means

Error "unknown combine_fn `{}`." thrown in tensorflow/models.

Source

Thrown at official/vision/modeling/decoders/nasfpn.py:286

      node0 = feats[input0]
      node0_level = feat_levels[input0]
      num_output_connections[input0] += 1
      node0 = self._resample_feature_map(node0, node0_level, new_level)
      node1 = feats[input1]
      node1_level = feat_levels[input1]
      num_output_connections[input1] += 1
      node1 = self._resample_feature_map(node1, node1_level, new_level)

      # Combine node0 and node1 to create new feat.
      if block_spec.combine_fn == 'sum':
        new_node = node0 + node1
      elif block_spec.combine_fn == 'attention':
        if node0_level >= node1_level:
          new_node = self._global_attention(node0, node1)
        else:
          new_node = self._global_attention(node1, node0)
      else:
        raise ValueError('unknown combine_fn `{}`.'
                         .format(block_spec.combine_fn))

      # Add intermediate nodes that do not have any connections to output.
      if block_spec.is_output:
        for j, (feat, feat_level, num_output) in enumerate(
            zip(feats, feat_levels, num_output_connections)):
          if num_output == 0 and feat_level == new_level:
            num_output_connections[j] += 1

            feat_ = self._resample_feature_map(feat, feat_level, new_level)
            new_node += feat_

      new_node = self._activation(new_node)
      new_node = self._conv_op(
          filters=self._config_dict['num_filters'],
          kernel_size=(3, 3),
          padding='same',
          **self._conv_kwargs)(new_node)

View on GitHub (pinned to e006f5f0d5)

Solutions

1. Use a supported combine_fn such as 'attention', 'sum', or 'concat' in the NASFPN config.
2. Check the combine_fn value for typos.

When it happens

Trigger: Thrown at official/vision/modeling/decoders/nasfpn.py:286 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/1a9e10b611ff0038. Report an issue: GitHub.