%1 = "tf.BroadcastTo"(%arg1, %cst) : (tensor<5x1xf32>, tensor<2xi64>) -> tensor<5x7xf32>
  %2 = "tf.Add"(%0, %1) : (tensor<5x7xf32>, tensor<5x7xf32>) -> tensor<5x7xf32>
  func.return %2 : tensor<5x7xf32>
  // CHECK: %[[V0:.*]] = "tf.Add"(%arg0, %arg1) : (tensor<7xf32>, tensor<5x1xf32>) -> tensor<5x7xf32>
  // CHECK: %[[V0]] : tensor<5x7xf32>
}

// CHECK-LABEL: @broadcast_batch_matmul_v2_rhs

    "tf.Yield"(%barg1, %barg2) : (tensor<!tf_type.variant<tensor<?x1xf32>>>, tensor<!tf_type.variant<tensor<?x1xf32>>>) -> ()
  }) {is_stateless = false} : (tensor<!tf_type.variant<tensor<?x1xf32>>>, tensor<!tf_type.variant<tensor<?x1xf32>>>) -> (tensor<!tf_type.variant<tensor<?x1xf32>>>, tensor<!tf_type.variant<tensor<?x1xf32>>>)
  func.return
}

// CHECK: while_region_op_two_sep_args_empty_tensor_list

  %5 = "tf.TPUExecute"(%3#0, %1, %program#0) {_producer_name = "UNKNOWN", device = "/TPU:0"} : (tensor<1x1xf32>, tensor<1x1xf32>, tensor<3x!tf_type.string>) -> tensor<1x1xf32>
  %6 = "tf.TPUExecute"(%3#1, %1, %program#1) {_producer_name = "UNKNOWN", device = "/TPU:1"} : (tensor<1x1xf32>, tensor<1x1xf32>, tensor<3x!tf_type.string>) -> tensor<1x1xf32>

    %9 = stablehlo.convert %2 : (tensor<2x3xi8>) -> tensor<2x3xf32>
    %10 = stablehlo.dot_general %8, %9, contracting_dims = [1] x [0] : (tensor<1x2xf32>, tensor<2x3xf32>) -> tensor<1x3xf32>
    %11 = stablehlo.convert %3 : (tensor<1x3xi32>) -> tensor<1x3xf32>
    %12 = stablehlo.subtract %10, %11 : tensor<1x3xf32>  // q1 * q2 - z1 * q2
    %13 = stablehlo.multiply %12, %4 : tensor<1x3xf32>  // s1 * s2

func.func @QuantizedCatsAddRequantsTest(%arg0: tensor<1x1xf32>, %arg1: tensor<1x1xf32>, %arg2: tensor<1x1xf32>, %arg3: tensor<1x1xf32>) -> (tensor<1x4xf32>, tensor<1x3xf32>) {
  %0 = "quantfork.stats"(%arg0) {layerStats = dense<[-0.440728068, 0.189515018]> : tensor<2xf32>} : (tensor<1x1xf32>) -> tensor<1x1xf32>
  %1 = "quantfork.stats"(%arg1) {layerStats = dense<[-0.154693216, 0.26483655]> : tensor<2xf32>} : (tensor<1x1xf32>) -> tensor<1x1xf32>

// CHECK-NEXT:    return [[RES]], [[MATRES]] : tensor<1x1xf32>, tensor<1x1xf32>
//
module {
  func.func @serving_default(%arg0: tensor<1x3xf32>) -> (tensor<1x1xf32>, tensor<1x1xf32>) {
    %0 = "tf.VarHandleOp"() <{container = "", shared_name = "y"}> : () -> tensor<!tf_type.resource<tensor<3x1xf32>>>
    %2 = "tf.ReadVariableOp"(%0) : (tensor<!tf_type.resource<tensor<3x1xf32>>>) -> tensor<3x1xf32>

  %outputs_2 = "tf.add"(%outputs_0, %outputs_1): (tensor<1x1xf32>, tensor<1x1xf32>) -> tensor<1x1xf32>
  return %outputs_2 : tensor<1x1xf32>
}

// CHECK-LABEL: @_func
func.func private @_func(%arg0: tensor<1x3xf32>, %arg1: tensor<3x1xf32>) -> (tensor<1x1xf32>) {

# CHECK-SAME:             -> (tensor<5x5xf32> {{{.*}}})
# CHECK-SAME: attributes {{.*}} tf_saved_model.exported_names = ["key"]

# CHECK-NEXT: [[R0:%.*]] = "tf.ReadVariableOp"([[ARG0]]) {{{.*}}} : (tensor<!tf_type.resource<tensor<5x3xf32>>>) -> tensor<5x3xf32>
# CHECK-NEXT: [[R1:%.*]] = "tf.ReadVariableOp"([[ARG1]]) {{{.*}}} : (tensor<!tf_type.resource<tensor<3x5xf32>>>) -> tensor<3x5xf32>

  func.func @simple_chain(%arg0: tensor<1xf32>) -> tensor<*xf32> {
    // CHECK: %[[MUL:.*]] = "tf.Mul"{{.*}} (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    // CHECK: %[[ADD:.*]] = "tf.Add"(%[[MUL]], %[[MUL]]) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    // CHECK: return %[[ADD]] : tensor<1xf32>
    %0 = "tf.Mul"(%arg0, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<*xf32>

// CHECK-NEXT:    "tf.MatMul"(%arg0, [[TENSOR]]) : (tensor<1x3xf32>, tensor<3x1xf32>) -> tensor<1x1xf32>
// CHECK-NEXT:    "tf.IfrtCall"(%arg0, [[ARRAYKEY]]) <{program_id = 6515870160938153680 : i64, variable_arg_indices = [1 : i32]}> {__tpu_compile_metadata_text = "retvals { sharding { } }"} : (tensor<1x3xf32>, tensor<!tf_type.string>) -> tensor<1x1xf32>
// CHECK-NEXT:    return
//