// CHECK: (%[[ARG:.*]]: tensor<4xf32>) -> (tensor<4xf32>, tensor<4xf32>, tensor<4xf32>)
func.func @main(%arg0: tensor<4xf32>) -> tensor<4xf32> attributes {tf.entry_function = {inputs = "mul"}} {
  %0 = "tfl.pseudo_const" () {value = dense<1.0> : tensor<4xf32>} : () -> tensor<4xf32> loc("Const")
  %1 = "tfl.squared_difference"(%arg0, %0) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("squared_difference")

  // CHECK: %[[DIV:.*]] = tfl.div
  %3 = "tfl.div"(%2, %1) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("div")
  // CHECK: %[[EXP:.*]] = "tfl.exp"
  %4 = "tfl.exp"(%3) : (tensor<4xf32>) -> tensor<4xf32> loc("exp")
  // tfl.neg should be pruned
  // CHECK-NOT: "tfl.neg"
  %5 = "tfl.neg"(%4) : (tensor<4xf32>) -> tensor<4xf32> loc("neg")

  %0 = "tfl.pseudo_const" () {value = dense<1.0> : tensor<4xf32>} : () -> tensor<4xf32> loc("Const")
  %1 = "tfl.mul"(%arg0, %0) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("mul")
  // tf.MyCustomOp is the result of conversion to a Custom op
  %2 = "tf.MyCustomOp"(%1, %0) {fused_activation_function = "RELU", int_attr = 2 : i32}  : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("MyCustomOp")

  %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} {
    %1 = "tf_device.launch"() <{device = "TPU_REPLICATED_HOST_0"}> ({
      %2 = "tf.Identity"(%arg_r0) : (tensor<1xf32>) -> tensor<1xf32>
      tf_device.return %2 : tensor<1xf32>
    }) : () -> tensor<1xf32>

  %0 = "tfl.pseudo_const" () {value = dense<1.0> : tensor<4xf32>} : () -> tensor<4xf32> loc("Const")
  %1 = "tfl.mul"(%arg0, %0) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("mul0")
  %2 = "tfl.mul"(%1, %0) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("mul1")
  %3 = "tfl.exp"(%2)  : (tensor<4xf32>) -> tensor<4xf32> loc("exp")
  func.return %3 : tensor<4xf32>

  %2:2 = "tf.RecvTPUEmbeddingActivations"() {_tpu_embedding_layer = "call1", config = "\0A\0B\0C\0D"} : () -> (tensor<2x2xf32>, tensor<4x4xf32>)
  func.return
}

// -----

// CHECK: while_region_op
func.func @while_region_op() {
  %elem_shape = "tf.Const"() {value = dense<[-1, 1]> : tensor<2xi32>} : () -> tensor<2xi32>
  %size = "tf.Const"() {value = dense<10> : tensor<i32>} : () -> tensor<i32>
  // CHECK: TensorListReserve
  // CHECK: TensorListReserve
  %tl = "tf.TensorListReserve"(%elem_shape, %size) : (tensor<2xi32>, tensor<i32>) -> tensor<!tf_type.variant<tensor<?x1xf32>>>
  %while:1 = "tf.WhileRegion"(%tl) ({

  %0 = "tf.FakeQuantWithMinMaxVars"(%arg0, %cst, %cst_0) {
    device = "", narrow_range = false, num_bits = 8 : i64
  } : (tensor<3xf32>, tensor<f32>, tensor<f32>) -> tensor<3xf32>
  %1 = "tf.FakeQuantWithMinMaxVarsPerChannel"(%arg1, %cst_1, %cst_2) {
    device = "", narrow_range = true, num_bits = 8 : i64
  } : (tensor<4x3xf32>, tensor<3xf32>, tensor<3xf32>) -> tensor<4x3xf32>
  func.return %0, %1 : tensor<3xf32>, tensor<4x3xf32>
}

  %1 = "tfl.mul"(%arg0, %0) {fused_activation_function = "NONE"} : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("mul")
  // tf.div is the result of conversion to a Flex TF op
  %2 = "tf.Div"(%1, %0)  : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32> loc("div")
  %3 = "tfl.exp"(%2)  : (tensor<4xf32>) -> tensor<4xf32> loc("exp")
  func.return %3 : tensor<4xf32>

// CHECK-LABEL: func @main
func.func @main(%arg0: tensor<1xf32>) -> tensor<1xf32>
attributes  {tf.entry_function = {inputs = "input", outputs = "output_node"}} {
  %0 = arith.constant dense<2.000000e+00> : tensor<f32>
  %1 = arith.constant dense<1.000000e+00> : tensor<f32>
  %2 = "tf.AddV2"(%arg0, %1) {T = "tfdtype$DT_FLOAT", device = "", name = "StatefulPartitionedCall/add"} : (tensor<1xf32>, tensor<f32>) -> tensor<1xf32>