}

  // CHECK-LABEL: func @_func
  // CHECK-SAME: %[[ARG0:.*]]: tensor<?xi32>,
  // CHECK-SAME: %[[ARG1:.*]]: tensor<?xi32> {mhlo.is_same_data_across_replicas = true}
  // CHECK-SAME: %[[ARG2:.*]]: tensor<?xi32>)
  func.func @_func(%arg0: tensor<?xi32>, %arg1: tensor<?xi32>, %arg2: tensor<?xi32>) -> tensor<?xi32> {
    %0 = "tf._D"(%arg0, %arg1) : (tensor<?xi32>, tensor<?xi32>) -> tensor<?xi32>
    func.return %0 : tensor<?xi32>

    // CHECK: %[[OUT:.*]], %[[MEAN:.*]], %[[VAR:.*]] = "mhlo.batch_norm_training"({{.*}}, %arg1, %arg2) <{epsilon = 1.000000e-03 : f32, feature_index = 3 : i64}> : (tensor<8x8x8x8xf32>, tensor<8xf32>, tensor<8xf32>) -> (tensor<8x8x8x8xf32>, tensor<8xf32>, tensor<8xf32>)

    func.return
}

func.func private @__reduce_func_2(%arg0: tensor<i64> {tf._user_specified_name = "args_0"}, %arg1: tensor<32xf32> {tf._user_specified_name = "args_1"}, %arg2: tensor<!tf_type.resource<tensor<64xf32>>>, %arg3: tensor<!tf_type.resource<tensor<128xf32>>>)
  -> (tensor<i64>) attributes {tf._tf_data_function = true, tf.signature.is_stateful} {

}

# CHECK-NO-SUBTYPE: func @main(%arg0: tensor<i32>, %arg1: tensor<!tf_type.variant>, %arg2: tensor<i8>, %arg3: tensor<f64>, %arg4: tensor<!tf_type.resource>) -> (tensor<i32>, tensor<!tf_type.variant>, tensor<i8>, tensor<f64>, tensor<!tf_type.resource>)

  // CHECK-DAG: ml_program.global public @vars.Variable_1([]) : tensor<!tf_type.string>
  func.func @lowers_string_ops(%arg0: tensor<128xi32>, %arg1: tensor<128xi32>, %arg2: tensor<128x1xi32>, %arg3: tensor<128x90xi32>, %arg4: tensor<128x90xi32>, %arg5: tensor<128x90xi32>, %arg6: tensor<128x90x64xf32>, %arg7: tensor<128x90x64xf32>) -> tensor<!tf_type.string> {
    // CHECK: %[[v0:.*]] = ml_program.global_load @vars.Variable_1 : tensor<!tf_type.string>

// CHECK: func private @tf.IfRegion_else(%arg0: tensor<i1>, %arg1: tensor<2xf32>, %arg2: tensor<2xf32>, %arg3: tensor<2xf32>) -> tensor<2xf32>
// CHECK-NEXT: "tf.Select"(%arg0, %arg2, %arg3)
// CHECK: func private @tf.IfRegion_then(%arg0: tensor<i1>, %arg1: tensor<2xf32>, %arg2: tensor<2xf32>, %arg3: tensor<2xf32>) -> tensor<2xf32>
// CHECK-NEXT: "tf.Select"(%arg0, %arg1, %arg2)
func.func @testIfCondition(%arg0: tensor<i1>, %arg1: tensor<2xf32>) -> tensor<2xf32> {

  func.return %4 : tensor<128x128x!quant.uniform<u8:f32, 0.2:125>>

// CHECK: %4 = "tfl.custom_tf"(%arg0, %arg1, %arg2, %arg3) ({
// CHECK-NEXT: ^bb0(%arg4: tensor<128x128xf32>, %arg5: tensor<1xf32>, %arg6: tensor<1xf32>, %arg7: tensor<1xi32>):

// CHECK-LABEL: func @FusedBatchNormRelu
func.func @FusedBatchNormRelu(%arg0: tensor<8x8x8x8xf32>, %arg1: tensor<8xf32>, %arg2: tensor<8xf32>, %arg3: tensor<8xf32>, %arg4: tensor<8xf32>) -> (tensor<8x8x8x8xf32>) {
// CHECK-NEXT: %[[Y:[a-z0-9]*]], {{.*}}_FusedBatchNormEx
// CHECK-NEXT: return %[[Y]]

// CHECK-LABEL: func @__forward_call_369
// CHECK-SAME: ([[in_chain:%.*]]: !tfrt.chain
// CHECK-SAME: [[arg1:%.*]]: !tfrt_fallback.tf_tensor {tf._user_specified_name = "inputs"},
// CHECK-SAME: [[arg2:%.*]]: !tfrt_fallback.tf_tensor, [[arg3:%.*]]: !tfrt_fallback.tf_tensor, [[arg4:%.*]]: !tfrt_fallback.tf_tensor, [[arg5:%.*]]: !tfrt_fallback.tf_tensor)
// CHECK-SAME: -> (!tfrt.chain
// CHECK: [[o1:%.*]] = tfrt_fallback_async.const_dense_tensor

// CHECK-SAME: (%[[ARG0:.*]]: tensor<f32>, %[[ARG1:.*]]: tensor<f32>, %[[ARG2:.*]]: tensor<*xf32>)
func.func @test_supported_lowering_of_tf_while_region(%arg0: tensor<f32>, %arg1: tensor<f32>, %arg2: tensor<*xf32>) -> (tensor<f32>){
  %0:2 = "tf.WhileRegion"(%arg0, %arg2) ({
  ^bb0(%arg3: tensor<f32>, %arg4: tensor<*xf32>):
    %1 = "tf.Identity"(%arg3) : (tensor<f32>) -> tensor<f32>