%0 = "tf.Const"() {device = "/device:CPU:0", value = dense<4.200000e+01> : tensor<f32>} : () -> tensor<f32>
  %1 = "tf.VarHandleOp"() {device = "/device:CPU:0", container = "", shape = #tf_type.shape<>, shared_name = "x"} : () -> tensor<!tf_type.resource<tensor<f32>>>
  "tf.AssignVariableOp"(%1, %0) {device = "/device:CPU:0"} : (tensor<!tf_type.resource<tensor<f32>>>, tensor<f32>) -> ()
  func.return

  %arg0: tensor<1x10x!quant.uniform<i8:f32, 0.1:-128>>,
  %arg1: tensor<1x5xf32>) -> (tensor<1x10x!quant.uniform<i8:f32, 0.2:42>>, tensor<1x5xf32>) {
  %0 = "tf.MyRequantize"(%arg0) : (tensor<1x10x!quant.uniform<i8:f32, 0.1:-128>>) -> tensor<1x10x!quant.uniform<i8:f32, 0.2:42>>
  %1 = "tf.Intermediate"(%arg1) : (tensor<1x5xf32>) -> tensor<1x5xf32>
  func.return %0, %1 : tensor<1x10x!quant.uniform<i8:f32, 0.2:42>>, tensor<1x5xf32>

module {
func.func @main(%arg0 : tensor<5xf32>, %arg1 : tensor<f32>, %arg2 : tensor<f32>) -> tensor<5xf32> {
  %0 = "tf.ClipByValue"(%arg0, %arg1, %arg2) : (tensor<5xf32>, tensor<f32>, tensor<f32>) -> tensor<5xf32>
  func.return %0 : tensor<5xf32>
}

  // CHECK: {{.*}} = corert.executeop.seq({{.*}}) "tf.ReadVariableOp"({{.*}}) {dtype = f32} : 1
  // CHECK: {{.*}} = corert.executeop({{.*}}) "tf.MatMul"({{.*}}) {T = f32, transpose_a = false, transpose_b = false} : 1
  %0 = "tf.ReadVariableOp"(%arg1) {device = "/device:GPU:0", dtype = f32} : (tensor<!tf_type.resource<tensor<1x3xf32>>>) -> tensor<1x3xf32>

  func.func private @composite_dequantize_uniform(%arg0: tensor<*xi8>) -> tensor<*xf32> {
    %cst = "tf.Const"() {value = dense<0.0787401571> : tensor<f32>} : () -> tensor<f32>
    %0 = "tf.Cast"(%arg0) {Truncate = false} : (tensor<*xi8>) -> tensor<*xf32>
    %1 = "tf.Mul"(%0, %cst) : (tensor<*xf32>, tensor<f32>) -> tensor<*xf32>
    return %1 : tensor<*xf32>
  }

}

// CHECK-LABEL: func @not_propagate_last_op

  // CHECK: "tf.Iterator"
  // CHECK-SAME: output_shapes{{.*}}200x28x28x1{{.*}}200x10
  // CHECK-SAME: output_types = [f32, f32]
  // CHECK-SAME: shared_name = "_iterator1"
  %0 = "tf.AnonymousIteratorV3"() {output_shapes = [
    #tf_type.shape<200x28x28x1>,
    #tf_type.shape<200x10>], output_types = [f32, f32]} : () -> tensor<!tf_type.resource>
  // CHECK: "tf.Iterator"
  // CHECK-SAME: shared_name = "_iterator2"

// CHECK-LABEL: @main

    %cst = stablehlo.constant dense<3.000000e-01> : tensor<2x3xf32>
    %0 = "quantfork.qcast"(%cst) : (tensor<2x3xf32>) -> tensor<2x3x!quant.uniform<i8:f32, 6.000000e-03>>
    %1 = "quantfork.dcast"(%0) : (tensor<2x3x!quant.uniform<i8:f32, 6.000000e-03>>) -> tensor<2x3xf32>

//===--------------------------------------------------------------------===//

// CHECK-LABEL: func @testPwStreamResults
func.func @testPwStreamResults(%arg0: tensor<f32>, %arg1: tensor<f32>) {
  "tf.PwStreamResults"(%arg0, %arg1) {names = ["foo", "bar"]} : (tensor<f32>, tensor<f32>) -> ()
  return
}

// -----
// CHECK-LABEL: func @test_ifrt_call
func.func @test_ifrt_call(%arg0: tensor<?xf32>, %arg1: tensor<?xf32>) {

      %launch:2 = "tf_device.launch"() ({
        %b:2 = "tf.opB"(%a) : (tensor<i1>) -> (tensor<i32>, tensor<f32>)
        tf_device.return %b#1, %b#0 : tensor<f32>, tensor<i32>
      }) {device = "CPU:0"} : () -> (tensor<f32>, tensor<i32>)
      %c = "tf.opC"() : () -> tensor<i1>
      tf_executor.yield %a, %launch#0, %launch#1, %c : tensor<i1>, tensor<f32>, tensor<i32>, tensor<i1>
    }
    tf_executor.fetch
  }
  func.return
}