const auto signature_def_map = metagraph_def.signature_def();
  const auto signature_def = signature_def_map.at("regress_x_to_y");

  const std::string input_name =
      signature_def.inputs().at(tensorflow::kRegressInputs).name();
  const std::string output_name =
      signature_def.outputs().at(tensorflow::kRegressOutputs).name();

        only supports float32 input.
    *   Add 5D support to `SLICE` op.
    *   TFLite Supports SingatureDef:
        *   TFLiteConverter exports models with SignatureDef
        *   Interpreter supports getting a list of signatures and getting
            callable function for a given signaturedef.
    *   Add int8 support for `ReshapeV2`.
    *   Add experimental support for optimization with sparsity.