// CHECK-NEXT:  } ]
// CHECK-NEXT: }

module attributes {tf.versions = {bad_consumers = [], min_consumer = 12 : i32, producer = 554 : i32}, tf_saved_model.semantics} {
  func.func @main() -> (tensor<5xf32> {tf_saved_model.index_path = ["start_logits"]}) attributes {tf.entry_function = {control_outputs = "", inputs = "", outputs = "StatefulPartitionedCall:1"}, tf_saved_model.exported_names = ["serving_default"]} {

  func.func @conv_with_filter_larger_than_1GB(%arg0: tensor<1x2240x2240x3xf32> {tf_saved_model.index_path = ["input"]}) -> (tensor<1x2240x1120x512xf32> {tf_saved_model.index_path = ["output"]}) attributes {tf.entry_function = {control_outputs = "", inputs = "serving_default_input:0", outputs = "PartitionedCall:0"}, tf_saved_model.exported_names = ["serving_default"]} {

  def __init__(self):
    super(TestModule, self).__init__()
    self.model = mnist_model()

  # CHECK: func {{@[a-zA-Z_0-9]+}}(%arg0: tensor<1x28x28x1xf32> {tf._user_specified_name = "x", tf_saved_model.index_path = [0]}
  # CHECK: attributes {{.*}} tf_saved_model.exported_names = ["my_predict"]
  @tf.function(input_signature=[
      tf.TensorSpec([1, 28, 28, 1], tf.float32),
  ])
  def my_predict(self, x):

module attributes {tf_saved_model.semantics, tf_saved_model.under_construction} {

  func.func @serving_default(%arg0: tensor<!tf_type.resource<tensor<100x50xf32>>> {tf.resource_name = "dense/kernel"}) -> (tensor<100x50xf32> {tf_saved_model.index_path = ["dense_2"]})
  attributes {tf.entry_function = {control_outputs = "", inputs = "", outputs = "dense_2/Add:0"}, tf_saved_model.exported_names = ["serving_default"]} {

  func.func @serving_default(%arg0: tensor<!tf_type.resource<tensor<100x50xf32>>> {tf.resource_name = "dense/kernel"}, %arg1: tensor<!tf_type.resource<tensor<50xf32>>> {tf.resource_name = "dense/bias"}) -> (tensor<100x50xf32> {tf_saved_model.index_path = ["dense_2"]})
  attributes {tf.entry_function = {control_outputs = "", inputs = "", outputs = "dense_2/Add:0"}, tf_saved_model.exported_names = ["serving_default"]} {

module attributes {tf.versions = {bad_consumers = [], min_consumer = 12 : i32, producer = 1269 : i32}, tf_saved_model.semantics} {
  func.func @depthwise_conv(%arg0: tensor<1x3x4x3xf32> {tf_saved_model.index_path = ["input_tensor"]}) -> (tensor<1x2x2x3xf32> {tf_saved_model.index_path = ["output"]}) attributes {tf.entry_function = {control_outputs = "", inputs = "serving_default_input_tensor:0", outputs = "PartitionedCall:0"}, tf_saved_model.exported_names = ["serving_default"]} {

}

// -----

module attributes {tf_saved_model.semantics} {
  func.func @conv_with_filter_larger_than_1MB(%arg0: tensor<1x224x224x3xf32> {tf_saved_model.index_path = ["input"]}) -> (tensor<1x224x112x512xf32> {tf_saved_model.index_path = ["output"]}) attributes {tf.entry_function = {control_outputs = "", inputs = "serving_default_input:0", outputs = "PartitionedCall:0"}, tf_saved_model.exported_names = ["serving_default"]} {

}

// CHECK-LABEL: func @set_readonly_stateless
func.func @set_readonly_stateless(%arg: tensor<i32> {tf_saved_model.index_path = ["arg"]}, %cond: tensor<i1> {tf_saved_model.index_path = ["cond"]}) -> (tensor<i32> {tf_saved_model.index_path = ["output"]}) attributes {tf_saved_model.exported_names = ["main"]} {

    return
  }

// CHECK: func.func @init_func_restore_op
// Check that an argument ("__tf_file_prefix") is created.
// CHECK-SAME: %[[ARG_0:.*]]: tensor<!tf_type.string> {tf_saved_model.index_path = ["__tf_file_prefix"]}

// Original `AssignVariableOp(VarHandleOp, Const)` pattern persists.
// CHECK-DAG: %[[CST_0:.*]] = "tf.Const"() {{.*value = dense<1.000000e\+00> : tensor<2xf32>.*}}

  EXPECT_TRUE(MessageDifferencer::Equals(method, Method::default_instance()));
}
constexpr absl::string_view kModuleDotWeightOnlyPtq = R"mlir(
  module {
    func.func @main(%arg0: tensor<?x2xf32> {tf_saved_model.index_path = ["input_tensor"]}) -> (tensor<?x2xf32>) {
      %0 = stablehlo.constant dense<[-0.211145893, -0.708605706]> : tensor<2xf32>