}

  Status AddInput(AbstractTensorHandle* input) override {
    GraphTensor* t = dyn_cast<GraphTensor>(input);
    if (!t) {
      return tensorflow::errors::InvalidArgument(
          "Unable to cast input to GraphTensor");
    }
    TF_AddInput(op_.get(), t->output_);
    return absl::OkStatus();
  }
  Status AddInputList(absl::Span<AbstractTensorHandle* const> inputs) override {

  // CHECK-DAG: %[[INPUT:.*]] = mhlo.bitcast_convert %arg0 : (tensor<3x2xi32>) -> tensor<3x2x!quant.uniform<i32:f32:1, {2.000000e+00:4,2.000000e+00:4}>>
  // CHECK-DAG: %[[MAX:.*]] = mhlo.bitcast_convert %arg1 : (tensor<3x2xi32>) -> tensor<3x2x!quant.uniform<i32:f32:1, {2.000000e+00:4,2.000000e+00:4}>>
  // CHECK-DAG: %[[INPUT_1:.*]] = chlo.broadcast_maximum

  func.func @main(%arg0: tensor<*x!tf_type.resource> {tf._user_specified_name = "input_1", tf.device = "/job:localhost/replica:0/task:0/device:CPU:0"}) attributes {allow_soft_placement = true, tf.entry_function = {control_outputs = "while,image_sample/write_summary/summary_cond", inputs = "image_sample_write_summary_summary_cond_input_1", outputs = ""}} {
    tf_executor.graph {

//
// void identity_shape_fn(TF_ShapeInferenceContext* ctx, TF_Status* status) {
//   TF_ShapeHandle* input = TF_NewShapeHandle();
//   TF_ShapeInferenceContextGetInput(ctx, 0, input, status);
//   if (TF_GetCode(status) == TF_OK) {
//     TF_ShapeInferenceContextSetOutput(ctx, 0, input, status);
//   }
//   TF_DeleteShapeHandle(input);
// }
//
// The following code registers the inference function with TensorFlow:
//

                    def input = inputArtifact.get().asFile
                    def output = outputs.file(input.name + ".blue")
                    assert output.parentFile.directory && output.parentFile.list().length == 0
                    println "Transforming \${input.name} to \${output.name}"
                    println "Input exists: \${input.exists()}"