// CHECK-DAG:           %[[VAL_3:.*]] = arith.constant dense<32> : tensor<2xi32>
// CHECK-DAG:           %[[VAL_4:.*]] = "tfl.resize_bilinear"(%[[VAL_2]], %[[VAL_3]]) <{align_corners = false, half_pixel_centers = true}> : (tensor<1x16x16x64xf32>, tensor<2xi32>) -> tensor<1x32x32x64xf32>
// CHECK-DAG:           %[[VAL_5:.*]] = arith.constant dense<[0, 3, 1, 2]> : tensor<4xi32>

  // since different model developers use RELU1 in different ways. Never
  // create another op called RELU1.
  RELU_N1_TO_1 = 20,
  RELU6 = 21,
  RESHAPE = 22,
  RESIZE_BILINEAR = 23,
  RNN = 24,
  SOFTMAX = 25,
  SPACE_TO_DEPTH = 26,
  SVDF = 27,
  TANH = 28,
  CONCAT_EMBEDDINGS = 29,
  SKIP_GRAM = 30,
  CALL = 31,
  CUSTOM = 32,
  EMBEDDING_LOOKUP_SPARSE = 33,

  // since different model developers use RELU1 in different ways. Never
  // create another op called RELU1.
  RELU_N1_TO_1 = 20,
  RELU6 = 21,
  RESHAPE = 22,
  RESIZE_BILINEAR = 23,
  RNN = 24,
  SOFTMAX = 25,
  SPACE_TO_DEPTH = 26,
  SVDF = 27,
  TANH = 28,
  CONCAT_EMBEDDINGS = 29,
  SKIP_GRAM = 30,
  CALL = 31,
  CUSTOM = 32,
  EMBEDDING_LOOKUP_SPARSE = 33,

// -----

// CHECK-LABEL: testResizeBilinear
func.func @testResizeBilinear(%arg0 : tensor<1x100x100x3xf32>, %arg1 : tensor<4xi32>) -> tensor<?xf32> {
  // CHECK: "tfl.resize_bilinear"(%arg0, %arg1) <{align_corners = false, half_pixel_centers = false}>
  %0 = "tfl.resize_bilinear"(%arg0, %arg1) {align_corners = false, half_pixel_centers = false} : (tensor<1x100x100x3xf32>, tensor<4xi32>) -> tensor<?xf32>
  func.return %0 : tensor<?xf32>
}

func.func @resize_with_bilinear(%arg0: tensor<1x100x100x3xf32>, %arg1: tensor<4xi32>) -> tensor<?xf32> {
  %0 = "tf.ResizeBilinear"(%arg0, %arg1) {align_corners = true} : (tensor<1x100x100x3xf32>, tensor<4xi32>) -> tensor<?xf32>
  func.return %0 : tensor<?xf32>
  // CHECK-LABEL: resize_with_bilinear
  // CHECK: "tfl.resize_bilinear"(%arg0, %arg1) <{align_corners = true, half_pixel_centers = false}> : (tensor<1x100x100x3xf32>, tensor<4xi32>) -> tensor<?xf32>

      // CHECK: %{{.*}} = stablehlo.custom_call @tf.ResizeBilinear(%arg0, %{{.*}}) {align_corners = false, device = "", half_pixel_centers = true} : (tensor<1x32x32x128xf32>, tensor<2xi32>) -> tensor<1x64x64x128xf32>

func.func @mhlo_custom_call_does_not_legalize_tf_function(%arg0: tensor<1x4xf32>) -> tensor<1x8xf32> {
  %0 = mhlo.custom_call @tf.ResizeBilinear(%arg0) {
    backend_config = "this_is_a_test_string"
  } : (tensor<1x4xf32>) -> (tensor<1x8xf32>)
  func.return %0 : tensor<1x8xf32>

  //       CHECK: %0 = mhlo.custom_call @tf.ResizeBilinear(%arg0) {
  //  CHECK-SAME:   backend_config = "this_is_a_test_string"

  Option<bool> skip_quantization_ops_{
      *this, "skip-quantization-ops",
      ::llvm::cl::desc("Skip quantization ops")};

  Option<bool> skip_resize_{
      *this, "skip-resize",
      ::llvm::cl::desc("Skip tf.ResizeBilinear and tf.ResizeNearestNeighbor")};

  Option<bool> skip_partitioned_calls_{
      *this, "skip-partitioned-calls",
      ::llvm::cl::desc(
          "Skip tf.StatefulPartitionedCall and tf.PartitionedCall")};
};

  let hasOptions = 1;
}

def TFL_ResizeBilinearOp: TFL_Op<"resize_bilinear", [
    Pure,
    QuantizableResult,
    PredOpTrait<"input and output must have same element type",
      TFL_TCresVTEtIsSameAsOp<0, 0>>,
    TFL_OperandHasRank<0, 4>,
    TFL_OperandHasRank<1, 1>,
    SameOperandsAndResultsScale]> {
  let summary = "ResizeBilinear Op";

  let description = [{

                      llvm::cl::Optional, llvm::cl::init(false));

// NOLINTNEXTLINE
opt<bool> skip_resize(
    "skip-resize",
    llvm::cl::desc(
        "Skip converting tf.ResizeBilinear and tf.ResizeNearestNeighbor ops."),
    llvm::cl::Optional, llvm::cl::init(true));

// NOLINTNEXTLINE
opt<bool> smuggle_disallowed_ops(
    "smuggle-disallowed-ops",