}

func.func private @tpu_func_0_optim0(%arg0: tensor<1x3x4x3xf32>) -> tensor<1x3x2x2xf32> attributes {tf._original_func_name = "tpu_func_0_optim"} {
  %cst = "tf.Const"() {device = "", value = dense_resource<__elided__> : tensor<2x3x3x2xbf16>} : () -> tensor<2x3x3x2xbf16>
  %cst_0 = "tf.Const"() {device = "", value = dense<[0, 3, 1, 2]> : tensor<4xi32>} : () -> tensor<4xi32>

func.func @NotQuantizeNoneType() -> none {
  %cst = "tfl.no_value"() {value = unit} : () -> none
  func.return %cst : none

// CHECK-NEXT:  %[[cst:.*]] = "tfl.no_value"() <{value}> : () -> none
// CHECK-NEXT:  return %[[cst]]
}

// CHECK-LABEL: QuantizeZeroSplat
func.func @QuantizeZeroSplat() -> tensor<2x3xf32> {
  %cst = arith.constant dense<0.0> : tensor<2x3xf32>
  func.return %cst : tensor<2x3xf32>

      tf_saved_model.initializer_type = "restore_op"} {
    %cst_0 = "tf.Const"() {value = dense<1.000000e+00> : tensor<2xf32>} : () -> tensor<2xf32>
    %var_0 = "tf.VarHandleOp"() {shared_name = "var_0"} : () -> tensor<!tf_type.resource<tensor<2xf32>>>
    "tf.AssignVariableOp"(%var_0, %cst_0) : (tensor<!tf_type.resource<tensor<2xf32>>>, tensor<2xf32>) -> ()
    return
  }

// CHECK-LABEL: avg_pool2d_3
// CHECK: %cst = arith.constant dense<[0, 2, 3, 1]> : tensor<4xi32>
// CHECK: %0 = "tfl.transpose"(%arg0, %cst) : (tensor<1x1x1x8xf32>, tensor<4xi32>) -> tensor<1x1x8x1xf32>
// CHECK{LITERAL}: %cst_0 = arith.constant dense<[[0, 0], [0, 0], [1, 1], [0, 0]]> : tensor<4x2xi32>
// CHECK: %1 = "tfl.pad"(%0, %cst_0) : (tensor<1x1x8x1xf32>, tensor<4x2xi32>) -> tensor<1x1x10x1xf32>

    %cst = "tf.Const"() <{value = dense<[0, 1]> : tensor<2xi32>}> {device = ""} : () -> tensor<2xi32>
    %cst_0 = "tf.Const"() <{value = dense<1.000000e+01> : tensor<f32>}> {device = ""} : () -> tensor<f32>
    %0 = "tf.Sum"(%arg0, %cst) <{keep_dims = false}> {device = ""} : (tensor<1x4xf32>, tensor<2xi32>) -> tensor<f32>
    %1 = "tf.Greater"(%0, %cst_0) {device = ""} : (tensor<f32>, tensor<f32>) -> tensor<i1>

// -----

module {
  func.func private @conv_with_bias_and_relu(%arg0: tensor<1x3x4x3xf32>) -> tensor<*xf32> {
    %cst = "tf.Const"() {device = "", value = dense<[7.11401462, 7.05456924]> : tensor<2xf32>} : () -> tensor<2xf32>

// CHECK-NOT: tf.ReadVariableOp
  %0 = "tf.ReadVariableOp"(%arg0) : (tensor<*x!tf_type.resource<tensor<1xf32>>>) -> tensor<1xf32>
  %cst = "tf.Const"() <{value = dense<0.000000e+00> : tensor<f32>}> : () -> tensor<f32>
  %cst_0 = "tf.Const"() <{value = dense<1> : tensor<1xi64>}> : () -> tensor<1xi64>
  %fill = "tf.Fill"(%cst_0, %cst) : (tensor<1xi64>, tensor<f32>) -> tensor<1xf32>
  tf_device.replicate([%0, %fill] as %arg_r0: tensor<1xf32>) {n = 2 : i32} {

    %cst = arith.constant dense<0.000000e+00> : tensor<5xf32>
    %cst_0 = arith.constant dense<1.0> : tensor<5x3xf32>
    %0 = "tfl.fully_connected"(%arg0, %cst_0, %cst) {fused_activation_function = "NONE", keep_num_dims = false, weights_format = "DEFAULT"} : (tensor<?x3xf32>, tensor<5x3xf32>, tensor<5xf32>) -> tensor<?x5xf32>

    %cst = "tf.Const"() {value = dense<3.00000000e-1> : tensor<2x3xf32>} : () -> tensor<2x3xf32>
    %cst_0 = "tf.Const"() {value = dense<4.00000000e-1> : tensor<1x3xf32>} : () -> tensor<1x3xf32>
    %0 = "quantfork.stats"(%arg0) {layerStats = dense<[6.00000000e-6, 9.00000000e-1]> : tensor<2xf32>} : (tensor<1x2xf32>) -> tensor<1x2xf32>

    %cst = "tf.Const"() {value = dense<6> : tensor<1xi32>} : () -> tensor<1xi32>
    %cst_0 = "tf.Const"() {value = dense<2> : tensor<1xi32>} : () -> tensor<1xi32>