Convert model to full int after quantization aware training

The Quantization aware training in Keras example mentions the following after performing quantization aware training (here):

After this, you have an actually quantized model with int8 weights and uint8 activations.

However, when the model is converted to TF Lite using the snippet of code from the example the resulting model is not a full integer model. For instance, these are what interpreter.get_input_details() and interpreter.get_output_details() return, respectively.

[{'name': 'input_1',
  'index': 0,
  'shape': array([ 1, 28, 28], dtype=int32),
  'shape_signature': array([-1, 28, 28], dtype=int32),
  'dtype': numpy.float32,
  'quantization': (0.0, 0),
  'quantization_parameters': {'scales': array([], dtype=float32),
   'zero_points': array([], dtype=int32),
   'quantized_dimension': 0},
  'sparsity_parameters': {}}]


[{'name': 'Identity',
  'index': 19,
  'shape': array([ 1, 10], dtype=int32),
  'shape_signature': array([-1, 10], dtype=int32),
  'dtype': numpy.float32,
  'quantization': (0.0, 0),
  'quantization_parameters': {'scales': array([], dtype=float32),
   'zero_points': array([], dtype=int32),
   'quantized_dimension': 0},
  'sparsity_parameters': {}}]

I've tried to convert that model to a full integer model using the standard TF Lite procedure (setting converter.target_spec.supported_ops, converter.inference_input_type, converter.inference_output_type, and converter.representative_dataset), however, the following error is raised RuntimeError: Quantization not yet supported for op: 'DEQUANTIZE'.

Is it possible to convert a model trained with quantization aware training to a full integer model?

Thanks,
Gustavo