// Casts the `target` type to a quantized type by using the quantization
// parameters from the type in the `source` type attribute.
// Examples:
//   f32 -> !quant.uniform<i8:f32, 1.0>
//   tensor<4xf32> -> tensor<4x!quant.uniform<i8:f32, 1.0>>
// The result is wrapped by a type attribute. Returns nullptr if the cast
// isn't valid.
//

   * @since 7.0 (source-compatible since 2.0)
   */
  // This the best we could do to get copyOfEnumSet to compile in the mainline.
  // The suppression also covers the cast to E[], discussed below.
  // In the backport, we don't have those cases and thus don't need this suppression.
  // We keep it to minimize diffs.
  @SuppressWarnings("unchecked")
  public static <E> ImmutableSet<E> copyOf(Collection<? extends E> elements) {

      Comparator<? super E> comparator, Iterable<? extends E> elements) {
    if (elements instanceof ImmutableSortedMultiset) {
      @SuppressWarnings("unchecked") // immutable collections are always safe for covariant casts
      ImmutableSortedMultiset<E> multiset = (ImmutableSortedMultiset<E>) elements;
      if (comparator.equals(multiset.comparator())) {
        if (multiset.isPartialView()) {

#include "tensorflow/core/framework/node_def_util.h"
#include "tensorflow/core/framework/rendezvous.h"
#include "tensorflow/core/framework/tensor_shape.pb.h"
#include "tensorflow/core/framework/types.h"
#include "tensorflow/core/platform/casts.h"
#include "tensorflow/core/platform/errors.h"
#include "tensorflow/core/platform/platform.h"
#include "tensorflow/core/platform/status.h"
#include "tensorflow/core/profiler/lib/traceme.h"

    if (!mlir::cast<ShapedType>(lhs.getType()).hasStaticShape() ||
        !mlir::cast<ShapedType>(rhs.getType()).hasStaticShape() ||
        !mlir::cast<ShapedType>(cond.getType()).hasStaticShape()) {
      return rewriteOpWithDynamicInput(op, rewriter);
    }

    auto lhs_shape = mlir::cast<ShapedType>(lhs.getType()).getShape();
    auto rhs_shape = mlir::cast<ShapedType>(rhs.getType()).getShape();

  auto input_args_have_static_shape = [&]() -> bool {
    return mlir::cast<TensorType>(input_tensor.getType()).hasStaticShape() &&
           mlir::cast<TensorType>(kernel_tensor.getType()).hasStaticShape() &&
           mlir::cast<TensorType>(window_strides.getType()).hasStaticShape() &&
           mlir::cast<TensorType>(padding.getType()).hasStaticShape() &&
           mlir::cast<TensorType>(lhs_dilation.getType()).hasStaticShape() &&

  }

  // Get the tensor types.
  const auto input_type = input_tensor.getType().cast<ShapedType>();
  const auto weights_type = weights_tensor.getType().cast<ShapedType>();
  const auto bias_type =
      has_bias ? bias_tensor.getType().cast<ShapedType>() : ShapedType{};

  const auto output_type = getType(0).cast<ShapedType>();

  // Folding only implemented for float tensors.

    auto op = cast<ResizeNearestNeighborOp>(src_op);
    auto loc = op.getLoc();
    auto result_ty = mlir::cast<ShapedType>(op.getType());

    auto input = op.getImages();
    auto input_ty = mlir::cast<ShapedType>(input.getType());
    auto input_element_ty = input_ty.getElementType();
    auto out_size = op.getSize();
    auto out_size_ty = mlir::cast<ShapedType>(out_size.getType());

    Operation* filter_op = filter_value.getDefiningOp();
    auto filter_uniform_quantized_type =
        GetElementType(filter_value).cast<UniformQuantizedPerAxisType>();
    auto filter_constant_value_attr = cast<DenseIntElementsAttr>(
        cast<stablehlo::ConstantOp>(filter_value.getDefiningOp()).getValue());
    const DenseIntElementsAttr new_filter_value_attr =

  }
}

LogicalResult SpaceToBatchNDOp::verify() {
  SpaceToBatchNDOp op = *this;
  const auto input_type = op.getInput().getType().cast<TensorType>();
  const auto block_shape_type = op.getBlockShape().getType().cast<TensorType>();
  const auto paddings_type = op.getPaddings().getType().cast<TensorType>();

  // Check that block_shape has rank 1.
  if (!IsOfRankOrUnranked(op.getBlockShape(), 1)) {