Does Crytpen support multiple outputs neural network？

[CHAOSV177]

in crypten/nn/onnx_converter.py there is:
def _to_crypten(onnx_model): """ Function that converts an onnx_model` to a CrypTen model.
"""

# create graph:
input_names, output_names = _get_input_output_names(onnx_model)
assert len(output_names) == 1, "Only one output per model supported."
crypten_model = module.Graph(input_names, output_names[0])`

it means only one output supported.
so it is not useful for many multiple outputs neural network.
I want to know how to change it so that Crypten can support multiple outputs neural network.