%0 = "tfl.less"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xi1>
  %1 = "tf.If"(%0, %arg0, %arg1) {else_branch = @cond_false, then_branch = @cond_true, is_stateless = false} : (tensor<1xi1>, tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  func.return %1 : tensor<1xf32>
}

  func.func private @if_then_branch(%arg0: tensor<*xi1>, %arg1: tensor<*xi64>, %arg2: tensor<*xi64>) -> tensor<*xi1> {
    func.return %arg0 : tensor<*xi1>
  }
  func.func private @if_else_branch(%arg0: tensor<*xi1>, %arg1: tensor<*xi64>, %arg2: tensor<*xi64>) -> tensor<*xi1> {
    func.return %arg0 : tensor<*xi1>
  }
  func.func @"tpu_subgraph"(%arg0: tensor<*xi64>) -> tensor<*xi1> {

  %3 = "tf.XlaSharding"(%arg1) { _XlaSharding = "\04\05\06", sharding = "\04\05\06" } : (tensor<*xi1>) -> tensor<*xi1>
  %4, %5 = "tf.A"(%2, %3) : (tensor<*xi1>, tensor<*xi1>) -> (tensor<*xi32>, tensor<*xi1>)
  %6 = "tf.XlaSharding"(%4) { _XlaSharding = "\0A\0B\0C", sharding = "\0A\0B\0C" } : (tensor<*xi32>) -> tensor<*xi32>

    %0 = "tf.opA"() : () -> (tensor<*xi1>)
    %1 = "tf.opB"() : () -> (tensor<*xi32>)
    tf_device.return %0, %1 : tensor<*xi1>, tensor<*xi32>
  },
  {
    %2 = "tf.opC"() : () -> (tensor<*xi1>)
    tf_device.return
  }) {} : () -> (tensor<*xi1>, tensor<*xi1>)
  func.return
}

// -----

// -----

// CHECK-LABEL: @if
func.func @if() {
  // CHECK-NOT: Optional
  %0 = builtin.unrealized_conversion_cast to tensor<5xi1>
  %1 = "tf.OptionalNone"() : () -> tensor<!tf_type.variant<tensor<f32>>>
  %2 = "tf.If"(%0, %1) <{else_branch = @false, is_stateless = false, then_branch = @true}>
      : (tensor<5xi1>, tensor<!tf_type.variant<tensor<f32>>>) -> (tensor<!tf_type.variant<tensor<f32>>>)
  return
}

    }
  }
}
versions {
  producer: 216
}

# NONE: func @main(%arg0: tensor<i1>, %arg1: tensor<1xi32>) -> (tensor<i1>, tensor<1xi32>)

  %0 = "tf.opA"() : () -> (tensor<*xi1>)
  %1 = "tf.opB"() : () -> (tensor<*xi1>)
  %2 = "tf.opC"() : () -> (tensor<*xi32>)
  %3 = "tf.opD"() : () -> (tensor<*xi32>)
  %4 = "tf.opE"() : () -> (tensor<*xf32>)
  %result:4 = tf_device.replicate([%0, %1] as %input0: tensor<*xi1>, [%2, %3] as %input1: tensor<*xi32>) {n = 2 : i32} {
    %5 = "tf.opF"(%input0, %4) : (tensor<*xi1>, tensor<*xf32>) -> (tensor<*xi1>)

      tf_executor.fetch %outputs_2 : tensor<*xi1>
    }
    return %0 : tensor<*xi1>
  }
  func.func private @image_sample_write_summary_summary_cond_false_710(%arg0: tensor<3x32x32x3xf32>, %arg1: tensor<*x!tf_type.resource>, %arg2: tensor<i64>) -> tensor<*xi1> attributes {tf._construction_context = "kEagerRuntime"} {
    %0 = tf_executor.graph {

  ^bb0(%arg0: tensor<i32>):
    %0 = "tf.Less"(%arg0, %arg0) {device = ""} : (tensor<i32>, tensor<i32>) -> tensor<*xi1>
    %1 = "tf.Less"(%arg0, %arg0) {device = ""} : (tensor<i32>, tensor<i32>) -> tensor<*xi1>
    %2 = "tf.LogicalAnd"(%0, %1) {device = ""} : (tensor<*xi1>, tensor<*xi1>) -> tensor<*xi1>
    "func.return"(%2) : (tensor<*xi1>) -> ()

    %6:2 = tf_executor.LoopCond %5#0 : (tensor<*xi1>) -> (tensor<*xi1>, !tf_executor.control) {device = ""} loc("while/LoopCond")