Targs = [tensor<32xf32>, tensor<64xf32>, tensor<8xf32>, tensor<2xf32>],
      Tresults = [tensor<32xf32>, tensor<64xf32>, tensor<8xf32>]}
      : (tensor<32xf32>, tensor<64xf32>, tensor<8xf32>, tensor<2xf32>, tensor<2x!tf_type.string>)
        -> (tensor<32xf32>, tensor<64xf32>, tensor<8xf32>)
    tf_device.return %0#0, %0#1, %0#2 : tensor<32xf32>, tensor<64xf32>, tensor<8xf32>

  // Use other output
  %3:6 = "tf.FusedBatchNormV3"( %2#0, %arg1, %arg2, %arg3, %arg4) {T = "tfdtype$DT_FLOAT", U = "tfdtype$DT_FLOAT", data_format = "NHWC", epsilon = 0.001 : f32, is_training = false} : (tensor<8x8x8x8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>) -> (tensor<8x8x8x8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>)

// CHECK: return %[[DEQUANTIZE]] : tensor<8xf32>
}

// CHECK-LABEL: fakeQuantNotFolded
func.func @fakeQuantNotFolded(tensor<8xf32>, tensor<f32>, tensor<f32>) -> (tensor<8xf32>) {
^bb0(%arg0: tensor<8xf32>, %arg3: tensor<f32>, %arg4: tensor<f32>):
  %1 = "tf.FakeQuantWithMinMaxVars"(%arg0, %arg3, %arg4) {num_bits = 3, narrow_range = false} : (tensor<8xf32>, tensor<f32>, tensor<f32>) -> tensor<8xf32>
  func.return %1 : tensor<8xf32>

  %0, %1, %2 ,%3, %4, %5 = "tf.FusedBatchNormV3"(%arg0, %arg1, %arg2, %arg3, %arg4) {T = "tfdtype$DT_BFLOAT16", U = "tfdtype$DT_BFLOAT16", data_format = "NHWC", epsilon = 0.001 : f32, is_training = false} : (tensor<8x8x8x8xbf16>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>, tensor<8xf32>) -> (tensor<8x8x8x8xbf16>, tensor<8xf32>, tensor<8xf32>,...

}

func.func @div(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
  %0 = "tf.Div"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  func.return %0: tensor<1xf32>

// CHECK-LABEL: div
// CHECK:  tfl.div %arg0, %arg1 {fused_activation_function = "NONE"} : tensor<1xf32>
// CHECK:  return
}

func.func @squaredDifferenceRelu(tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32> {

  tf_executor.graph {
    %outputs, %control = tf_executor.island wraps "tf.ReadVariableOp"(%arg0) {device = ""} : (tensor<*x!tf_type.resource<tensor<8xf32>>>) -> tensor<8xf32>

    %res:2 = tf_executor.Enter %arg0 frame "some/fra\"me" : (tensor<8xf32>) -> (tensor<*xf32>, !tf_executor.control)
    tf_executor.fetch %res#0 : tensor<*xf32>
  }
  func.return %result : tensor<*xf32>
}

// CHECK-LABEL: func @enter_parallel_iterations(%{{.*}}: tensor<*xf32>, %{{.*}}: i1) -> tensor<*xf32> {
func.func @enter_parallel_iterations(%arg0: tensor<*xf32>, %arg1: i1) -> tensor<*xf32> {

     %2 = "tf.ReadVariableOp"(%arg2) {device = ""} : (tensor<*x!tf_type.resource<tensor<8xf32>>>) -> tensor<8xf32>
     "tf.LoadTPUEmbeddingAdagradParameters"(%0, %1) {config = "", num_shards = 1 : i64, shard_id = 0 : i64, table_id = -1 : i64, table_name = "table1"} : (tensor<8xf32>, tensor<8xf32>) -> ()
     %3 = "tf.ReadVariableOp"(%arg3) {device = ""} : (tensor<*x!tf_type.resource<tensor<8xf32>>>) -> tensor<8xf32>

     %2 = "tf.ReadVariableOp"(%arg2) {device = ""} : (tensor<*x!tf_type.resource<tensor<8xf32>>>) -> tensor<8xf32>
     "tf.LoadTPUEmbeddingAdagradParameters"(%0, %1) {config = "", num_shards = 1 : i64, shard_id = 0 : i64, table_id = -1 : i64, table_name = "table1"} : (tensor<8xf32>, tensor<8xf32>) -> ()
     %3 = "tf.ReadVariableOp"(%arg3) {device = ""} : (tensor<*x!tf_type.resource<tensor<8xf32>>>) -> tensor<8xf32>

    %1:2 = tf_executor.island wraps "tf.StatefulPartitionedCall"(%0#0) {Tin = ["tfdtype$DT_RESOURCE"], Tout = ["tfdtype$DT_RESOURCE"], config = "", config_proto = "", executor_type = "", f = @foo} : (tensor<!tf_type.resource<tensor<8xf32>>>) -> tensor<!tf_type.resource<tensor<8xf32>>> loc("call_foo")
    tf_executor.fetch
  }
  func.return
}