return
  }
  // CHECK-LABEL: func private @foo.impl
  func.func private @foo.impl(%arg0: tensor<1xf32>, %arg1: tensor<2xf32>) -> (tensor<2xf32>, tensor<1xf32>) {
    return %arg1, %arg0 : tensor<2xf32>, tensor<1xf32>
  }

  // CHECK-DAG: %[[CST_0:.+]] = mhlo.constant dense<[1.000000e-01, 2.000000e-01]> : tensor<2xf32>
  // CHECK-DAG: %[[CST_1:.+]] = mhlo.constant dense<[3.000000e-01, 4.000000e-01]> : tensor<2xf32>
  // CHECK: %[[CST_BCAST:.+]] = "mhlo.broadcast_in_dim"(%[[CST_0]]) <{broadcast_dimensions = dense<3> : tensor<1xi64>}> : (tensor<2xf32>) -> tensor<1x1x3x2xf32>

//RUN: tf_tfl_translate --enable-stablehlo-conversion --input-mlir %s -o /tmp/temp.stablehlo; [ -f /tmp/temp.stablehlo ]


module {
func.func @main(%arg0: tensor<2xf32>) -> tensor<2xf32> {
  %0 = "tf.Sigmoid"(%arg0) : (tensor<2xf32>) -> tensor<2xf32>
  func.return %0 : tensor<2xf32>
}

    } : (tensor<2x2xf32>, tensor<2x2xf32>) -> tensor<2x2xf32>
    %2 = "quantfork.stats"(%1) {layerStats = dense<[0.000000e+00, 6.000000e+00]> : tensor<2xf32>} : (tensor<2x2xf32>) -> tensor<2x2xf32>
    return %2 : tensor<2x2xf32>
  }

  // CHECK-LABEL: composite_dot_general
  func.func private @composite_dot_general(%arg0: tensor<2x2xf32>, %arg1: tensor<2x2xf32>) -> tensor<2x2xf32> {

  %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.return
}

func.func private @__reduce_func_2(%arg0: tensor<i64> {tf._user_specified_name = "args_0"}, %arg1: tensor<32xf32> {tf._user_specified_name = "args_1"}, %arg2: tensor<!tf_type.resource<tensor<64xf32>>>, %arg3: tensor<!tf_type.resource<tensor<128xf32>>>)

    %a:2, %control1 = tf_executor.island wraps "tf.A"() {_class = ["loc:@class"]} : () -> (tensor<2xf32>, tensor<2xf32>)
    %s:2, %control2 = tf_executor.island wraps "tf.Split"(%c, %a#1) {num_split = 2 : i32} : (tensor<i32>, tensor<2xf32>) -> (tensor<1xf32>, tensor<1xf32>)
    tf_executor.fetch
  }
  func.return
}

// -----

// CHECK-LABEL: func @no_colocate_split_has_device

func.func @transpose_simple_1d() -> tensor<2xf32> {
  %0 = stablehlo.constant dense<[0.000000e+0, 1.000000e+0]> : tensor<2xf32>
  %1 = stablehlo.transpose %0, dims = [0] : (tensor<2xf32>) -> tensor<2xf32>
  return %1 : tensor<2xf32>
}
// CHECK-DAG: %[[CONST_0:.+]] = stablehlo.constant dense<[0.000000e+00, 1.000000e+00]> : tensor<2xf32>
// CHECK-NOT: transpose
// CHECK: return %[[CONST_0]] : tensor<2xf32>

// -----

  %1 = stablehlo.constant dense<[4.0, 6.0]> : tensor<2xf32>
  %2 = stablehlo.divide %0, %1 : tensor<2xf32>
  return %2 : tensor<2xf32>
}

// CHECK-LABEL: foldDivLHSSplat
// CHECK: stablehlo.constant dense<[5.000000e-01, 0.333333343]> : tensor<2xf32>

// -----

func.func @foldDivRHSSplat() -> tensor<2xf32> {
  %0 = stablehlo.constant dense<[4.0, 6.0]> : tensor<2xf32>
  %1 = stablehlo.constant dense<2.0> : tensor<2xf32>

  %1 = "quantfork.stats"(%0) {bitsNum = 8 : i64, layerStats = dense<[-2.0, 2.0]> : tensor<2xf32>, narrowRange = false} : (tensor<2x3xf32>) -> tensor<2x3xf32>
  %2 = stablehlo.convert %1 : (tensor<2x3xf32>) -> (tensor<2x3xf32>)
  func.return %2 : tensor<2x3xf32>
}

// -----

// CHECK-LABEL: func @dot_redundant_stats