tensorflow/models · error · ValueError

Currently only supports FPN.

Error message

Currently only supports FPN.

What it means

Error "Currently only supports FPN." thrown in tensorflow/models.

Source

Thrown at official/projects/qat/vision/modeling/factory.py:173

        model=model,
        **model.checkpoint_items)
    status = ckpt.read(original_checkpoint)
    status.expect_partial().assert_existing_objects_matched()

  scope_dict = {
      'L2': tf_keras.regularizers.l2,
      'BatchNormalizationWrapper': qat_nn_layers.BatchNormalizationWrapper,
  }
  with tfmot.quantization.keras.quantize_scope(scope_dict):
    annotated_backbone = tfmot.quantization.keras.quantize_annotate_model(
        model.backbone)
    optimized_backbone = tfmot.quantization.keras.quantize_apply(
        annotated_backbone,
        scheme=schemes.Default8BitQuantizeScheme())
    decoder = model.decoder
    if quantization.quantize_detection_decoder:
      if not isinstance(decoder, fpn.FPN):
        raise ValueError('Currently only supports FPN.')

      decoder = tf_keras.models.clone_model(
          decoder,
          clone_function=_clone_function_for_fpn,
      )
      decoder = tfmot.quantization.keras.quantize_apply(decoder)
      decoder = tfmot.quantization.keras.remove_input_range(decoder)

    head = model.head
    if quantization.quantize_detection_head:
      if not isinstance(head, dense_prediction_heads.RetinaNetHead):
        raise ValueError('Currently only supports RetinaNetHead.')
      head = (
          dense_prediction_heads_qat.RetinaNetHeadQuantized.from_config(
              head.get_config()))

  optimized_model = retinanet_model.RetinaNetModel(
      backbone=optimized_backbone,

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/qat/vision/modeling/factory.py:173 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/ed39c5581e8ebd42. Report an issue: GitHub.