Output* x, Output* y) {
    *x = Const<T>(scope_, x_data);
    switch (op_type) {
      case RESIZE_NEAREST:
        *y = ResizeNearestNeighbor(
            scope_, *x, y_shape,
            ResizeNearestNeighbor::AlignCorners(align_corners));
        return;
      case RESIZE_BILINEAR:
        *y = ResizeBilinear(scope_, *x, y_shape,

}

def TFL_ResizeNearestNeighborOp : TFL_Op<"resize_nearest_neighbor", [
    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 = "ResizeNearestNeighbor Op";

  let description = [{

      internal::ResizeNearestNeighborGrad::AlignCorners(align_corners)
          .HalfPixelCenters(half_pixel_centers)));
  grad_outputs->push_back(NoGradient());
  return scope.status();
}
REGISTER_GRADIENT_OP("ResizeNearestNeighbor", ResizeNearestNeighborGradHelper);

Status ResizeBilinearGradHelper(const Scope& scope, const Operation& op,
                                const std::vector<Output>& grad_inputs,

^bb0(%arg0: tensor<? x ? x ? x ? x !quant.uniform<u8:f32, 0.1>>, %arg1: tensor<? x i32>):
  %0 = "tfl.resize_nearest_neighbor"(%arg0, %arg1) { align_corners = false, half_pixel_centers = false } : (tensor<? x ? x ? x ? x !quant.uniform<u8:f32, 0.1>>, tensor<? x i32>) -> tensor<? x !quant.uniform<u8:f32, 0.1>>

                      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",

  // b/135640736: MatrixInverse performance issues.
  // b/111271662: MatrixSolve performance issues.
  // https://github.com/tensorflow/tensorflow/pull/31012:
  //    ResizeNearestNeighbor, ResizeBilinear, and ResizeBilinearGrad sometimes
  //    create convolutions too large for CuDNN to handle.
  return node.type_string() == "SelfAdjointEigV2" ||

  // CHECK: [[VAL5:%.+]] = "tf.GatherV2"([[VAL4]], [[VAL3]], [[VAL0]]) <{batch_dims = 0 : i64}>
  // CHECK: [[VAL6:%.+]] = "tf.Reshape"([[VAL5]], [[VAL1]])
  // CHECK: return [[VAL6]]
  %resize = "tf.ResizeNearestNeighbor"(%arg0, %shape) {align_corners = false, device = "", half_pixel_centers = false} : (tensor<1x7x7x1xi32>, tensor<2xi32>) -> tensor<1x3x3x1xi32>
  func.return %resize: tensor<1x3x3x1xi32>
}

    "LOGICAL_NOT",
    "UNPACK",
    "REDUCE_MIN",
    "FLOOR_DIV",
    "REDUCE_ANY",
    "SQUARE",
    "ZEROS_LIKE",
    "FILL",
    "FLOOR_MOD",
    "RANGE",
    "RESIZE_NEAREST_NEIGHBOR",
    "LEAKY_RELU",
    "SQUARED_DIFFERENCE",
    "MIRROR_PAD",
    "ABS",
    "SPLIT_V",
    "UNIQUE",
    "CEIL",
    "REVERSE_V2",
    "ADD_N",
    "GATHER_ND",

      Operation *op = rewriter.create(state);
      result = op->getResult(0);
    }
    rewriter.replaceOp(op, {result});
    return success();
  }
};

// Lowers ResizeNearestNeighbor to an indices computations with a gather along
// the combined spatial dimensions. Generating the indices along the
// width/height index could be used to gather along each of W and H dimension

      "RandomStandardNormal",
      "RandomUniform",
      "RandomUniformInt",
      "ReadVariableOp",
      "ReadVariableXlaSplitND",
      "ResizeBilinear",
      "ResizeBilinearGrad",
      "ResizeNearestNeighbor",
      "ResourceApplyAdaMax",
      "ResourceApplyAdadelta",
      "ResourceApplyAdagrad",
      "ResourceApplyAdagradDA",
      "ResourceApplyAdagradV2",
      "ResourceApplyAdam",