If you are a TensorFlow Python user, you can try this format by setting the
`experimental_image_format` option:

```
tf.savedmodel.save(
    model, path,
    options=tf.saved_model.SaveOptions(experimental_image_format=True)
)
```

When this option is enabled, exported SavedModels with proto size > 2GB will
automatically save with the new format (`.cpb` instead of `.pb`).

<!-- **Compatibility** -->

  root.get_c = def_function.function(lambda: root.c)
  signatures = root.get_c.get_concrete_function()
  save.save(root, non_chunked_dir, signatures=signatures,
            options=save_options.SaveOptions(experimental_image_format=False))


def generate_chunked_model(non_chunked_dir: str, chunked_dir: str):
  saved_model = loader_impl.parse_saved_model(non_chunked_dir)

    """Save a saved model with unsupported ops, and then load and convert it."""
    # saved the model
    test_model = TestModule()
    saved_model_path = '/tmp/test.saved_model'
    save_options = tf.saved_model.SaveOptions(save_debug_info=True)
    tf.saved_model.save(test_model, saved_model_path, options=save_options)

    # load the model and convert
    converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_path)

  if not module_ctor:
    print("Expected ModuleName to be one of:", MODULE_CTORS.keys())
    return 2
  os.makedirs(export_path)

  tf_module = module_ctor()
  if version == 2:
    options = save_options.SaveOptions(save_debug_info=True)
    saved_model.save(tf_module, export_path, options=options)
  else:
    builder = saved_model.builder.SavedModelBuilder(export_path)
    builder.add_meta_graph_and_variables(tf_module, ["serve"])

    if FLAGS.save_model_path:
      save_model_path = FLAGS.save_model_path
    else:
      save_model_path = tempfile.mkdtemp(suffix='.saved_model')
    save_options = tf.saved_model.SaveOptions(save_debug_info=show_debug_info)
    tf.saved_model.save(
        create_module_fn(), save_model_path, options=save_options
    )
    logging.info('Saved model to: %s', save_model_path)

          out = activation_fn(out)
        return {'output': out}

    model = ConvModel()
    save_options = None
    if has_func_alias:
      save_options = tensorflow.saved_model.SaveOptions(
          function_aliases={FUNC_ALIAS: model.conv2d}
      )
    saved_model_save.save(
        model,
        saved_model_path,
        signatures=model.conv2d.get_concrete_function(

        input_shape=(1, 3, 4, 3), filter_shape=(2, 3, 3, 2)
    )

    signatures = {
        'serving_default': model.conv.get_concrete_function(),
    }
    save_opts = save_options.SaveOptions(
        function_aliases={'conv_func': model.conv}
    )

    saved_model_save.save(
        model, self._input_saved_model_path, signatures, save_opts
    )

* `tf.train.CheckpointOptions`
    * Added `experimental_skip_slot_variables` (a boolean option) to skip
    restoring of optimizer slot variables in a checkpoint.

*   `tf.saved_model.SaveOptions`

    * `SaveOptions` now takes a new argument called
      `experimental_debug_stripper`. When enabled, this strips the debug nodes
      from both the node defs and the function defs of the graph. Note that