//CHECK-NEXT: return %0 : tensor<1x30x1xi32>
//CHECK-NEXT:}

func.func @power(%arg0: tensor<1x30x1xi32>, %arg1: tensor<1x30x1xi32>) -> tensor<1x30x1xi32> {
  %0 = "vhlo.power_v1" (%arg0, %arg1) : (tensor<1x30x1xi32>, tensor<1x30x1xi32>) -> tensor<1x30x1xi32>
  func.return %0 : tensor<1x30x1xi32>
}

//CHECK:func.func private @power(%arg0: tensor<1x30x1xi32>, %arg1: tensor<1x30x1xi32>) -> tensor<1x30x1xi32> {

  // CHECK:  %3 = mhlo.reshape %2 : (tensor<32xi32>) -> tensor<1x32x1xi32>
  // CHECK:  %cst = arith.constant dense<1> : tensor<1xi32>
  // CHECK:  %4 = "tfl.reduce_max"(%arg0, %cst) <{keep_dims = false}> : (tensor<1x32x1xf32>, tensor<1xi32>) -> tensor<1x1xf32>
  // CHECK:  %5 = "tfl.arg_max"(%arg0, %cst) : (tensor<1x32x1xf32>, tensor<1xi32>) -> tensor<1x1xi32>
  // CHECK:  return %4, %5 : tensor<1x1xf32>, tensor<1x1xi32>
}

    %11 = stablehlo.convert %2 : (tensor<1x1x1xi8>) -> tensor<1x1x1xi32>
    %12 = stablehlo.broadcast_in_dim %11, dims = [0, 1, 2] : (tensor<1x1x1xi32>) -> tensor<8x16x16xi32>
    %13 = stablehlo.subtract %10, %12 : tensor<8x16x16xi32>  // q1 - z1
    %14 = stablehlo.convert %9 : (tensor<8x16x4xi8>) -> tensor<8x16x4xi32>
    %15 = stablehlo.convert %4 : (tensor<1x1x1xi8>) -> tensor<1x1x1xi32>

  %cst = arith.constant dense<[1, -1, 4]> : tensor<3xi32>
  // CHECK: "tfl.reshape"(%arg0, %cst)
  %0 = "tfl.reshape"(%arg0, %cst) : (tensor<3x4xi32>, tensor<3xi32>) -> tensor<1x3x4xi32>
  func.return %0 : tensor<1x3x4xi32>
}

// -----

func.func @invalid_two_dynamic_dims_on_reshape(%arg0: tensor<3x4xi32>) -> tensor<1x3x4xi32> {

// Examples:
//   * If `xla_gather_op_output_type` == tensor<*xf32>, then it returns:
//     tensor<*xf32>.
//   * If `xla_gather_op_output_type` == tensor<3x5xi32> and `collapsed_dims` ==
//     {0}, then it returns: tensor<1x3x5xi32>.
//   * If `xla_gather_op_output_type` == tensor<3x5xf32> and `collapsed_dims` ==
//     {1, 3}, then it returns: tensor<3x1x5x1xf32>.
Type GetSliceOpOutputType(Type xla_gather_op_output_type,

  func.return %resize: tensor<1x3x3x1xi32>
}

// CHECK-LABEL: func @imag_resize_nearest_dyn_img
func.func @imag_resize_nearest_dyn_img(%arg0: tensor<1x?x?x1xi32>) -> tensor<1x3x3x1xi32> {
  %shape = "tf.Const"() {device = "", value = dense<3> : tensor<2xi32>} : () -> tensor<2xi32>

// CHECK-DAG:       %[[VAL_9:.*]] = arith.constant dense<1> : tensor<1xi32>
// CHECK:           %[[VAL_10:.*]] = "tf.Max"(%[[VAL_0]], %[[VAL_9]]) <{keep_dims = false}> : (tensor<1x32x1xf32>, tensor<1xi32>) -> tensor<1x1xf32>
// CHECK:           %[[VAL_11:.*]] = "tf.ArgMax"(%[[VAL_0]], %[[VAL_9]]) : (tensor<1x32x1xf32>, tensor<1xi32>) -> tensor<1x1xi32>

}

// -----

// CHECK-LABEL: func @selectV2BroadcastAll
func.func @selectV2BroadcastAll(%arg0: tensor<8x1x1xi1>, %arg1: tensor<1x8x1xi32>, %arg2: tensor<1x1x8xi32>) -> tensor<8x8x8xi32> {
  %0 = "tf.SelectV2"(%arg0, %arg1, %arg2) : (tensor<8x1x1xi1>, tensor<1x8x1xi32>, tensor<1x1x8xi32>) -> tensor<8x8x8xi32>
  func.return %0: tensor<8x8x8xi32>
}

// -----

// CHECK-LABEL: func @selectV2DynamicRanked