Int8 static quantization on Mediapipe SelfieMulticlass model requiring signature_key

[BarbaraCorrieroArm]

I am trying to quantize the .tflite model from mediapipe segmentation task SelfieMulticlass (256 x 256) using the recipe static i8 weight and i8 activations.
I have added this to recipe.py:

def static_wi8_ai8():
  """Returns a static quantization recipe with int8 weights and int8 activation."""
  return [
      dict({
          'regex': '.*',
          'operation': '*',
          'algorithm_key': 'min_max_uniform_quantize',
          'op_config': {
              'weight_tensor_config': {
                  'num_bits': 8,
                  'symmetric': True,
                  'granularity': 'CHANNELWISE',
                  'dtype': 'INT',
                  'block_size': 0,
              },
              'activation_tensor_config': {
                  'num_bits': 8,
                  'symmetric': True,
                  'granularity': 'CHANNELWISE',
                  'dtype': 'INT',
                  'block_size': 0,
              },
              'compute_precision': 'INTEGER',
              'explicit_dequantize': False,
              'skip_checks': False,
          },
      })
  ]  

and I am trying to use a calibration test dataset to calibrate.

    # Define the representative dataset generator
    def representative_data_gen():
        for _ in range(100):
            # Replace "input_29" with your input name
            yield {"default": np.random.rand(1, 256, 256, 3).astype(np.float32)}

    # Use a fallback signature key ("default") and pass our generator
    calibration_result = qt.calibrate({
        "None": representative_data_gen()    #tried "default", "input_29", "main", "serving_default", "Conv2D"
                                         # "" - Exception has occurred: ValueError None signature_key provided.
                                         # None - 'SignatureDef signature_key is None and model has {0} Signatures. '
                                         #'None is only allowed when the model has 1 SignatureDef'.
    })
    print("Calibration completed successfully.")

I keep getting errors on pretty much every signature key I have tried. The point is the model doesn't have one as I see when I try to print it:
=== Inputs:
name: input_29, shape: [ 1 256 256 3], dtype: <class 'numpy.float32'>

===Outputs:
name: Identity, shape: [ 1 256 256 6], dtype: <class 'numpy.float32'>

=== Signatures:
No signatures found in the model.
{}

I have also tried "None" but I get the error listed above in the comment.
I'm just wondering why we need a signature key. My understanding is that we need signature keys only when a model has multiple inputs/outputs and there is only one here.