auto splat_elements_attr =
        mlir::dyn_cast<SplatElementsAttr>(const_op.getValue());
    if (!splat_elements_attr) {
      return;
    }
    auto element_type = splat_elements_attr.getType().getElementType();
    if (!(element_type.isF32() || element_type.isF16() ||
          element_type.isInteger(1) || element_type.isInteger(32) ||
          element_type.isInteger(64))) {
      return;
    }

  if (!begin_attr || !size_attr) {
    return false;
  }

  auto begin_elem_ty = begin_attr.getType().getElementType();
  if (!begin_elem_ty.isInteger(32) && !begin_elem_ty.isInteger(64)) {
    return false;
  }
  auto size_elem_ty = size_attr.getType().getElementType();
  if (!size_elem_ty.isInteger(32) && !size_elem_ty.isInteger(64)) {
    return false;
  }

      func_.getFunctionType().getInput(0));
  if (!image_type || !image_type.getElementType().isF32() ||
      image_type.getRank() != 4) {
    return func_.emitWarning() << "Image should be a 4D float tensor";
  }

  auto flow_type = mlir::dyn_cast_or_null<RankedTensorType>(
      func_.getFunctionType().getInput(1));
  if (!flow_type || !flow_type.getElementType().isF32() ||
      flow_type.getRank() != 4) {

 * limitations under the License.
 */

package org.gradle.api.internal.provider;

public interface CollectionProviderInternal<T, C extends Iterable<T>> extends ProviderInternal<C> {
    Class<? extends T> getElementType();

    int size();

  auto transpose_shape =
      llvm::to_vector<8>(llvm::map_range(perm, [transpose_type](int32_t dim) {
        return transpose_type.getDimSize(dim);
      }));
  auto elem_type = transpose_type.getElementType();
  auto result_type = RankedTensorType::get(transpose_shape, elem_type);

  return builder->create<TF::TransposeOp>(location, result_type,
                                          value_to_transpose, perm_op);
}

  for (int64_t i = 0; i < permutation_array.size(); ++i) {
    transposed_shape[i] = input_type.getDimSize(permutation_array[i]);
  }
  auto transposed_type =
      RankedTensorType::get(transposed_shape, input_type.getElementType());
  DenseIntElementsAttr permutation = DenseIntElementsAttr::get(
      RankedTensorType::get(permutation_array.size(), rewriter.getI64Type()),
      permutation_array);
  return {permutation, transposed_type};
}

    private ModelSchema<E> elementTypeSchema;

    public CollectionSchema(ModelType<T> type, ModelType<E> elementType) {
        super(type);
        this.elementType = elementType;
    }

    public ModelType<E> getElementType() {
        return elementType;
    }

    public ModelSchema<E> getElementTypeSchema() {
        return elementTypeSchema;
    }

    @Override
    public String toString() {

bool IsExpressedTypeF32(QuantizedType quantized_type);

// Given a value, extract the `ElementType`.
// `value` should be a non-null `TensorType`.
inline Type GetElementType(const Value value) {
  return mlir::cast<TensorType>(value.getType()).getElementType();
}

// Returns true iff `type` is a uniform quantized type whose storage type is
// 8-bit integer and expressed type is f32.
bool IsI8F32UniformQuantizedType(Type type);

    Value iota = reduce_op.getInputs().back();
    if (!MatchIota(reduce_op.getDimensions(), iota)) return failure();

    // Match the reduction computation.
    const bool is_float = mlir::isa<FloatType>(operand_init.getElementType());
    if (failed(MatchReduceToArgMinMaxType1(reduce_op, is_float, is_argmax)) &&
        failed(MatchReduceToArgMinMaxType2(reduce_op, is_argmax)))
      return rewriter.notifyMatchFailure(

 public:
  using ConvertReduceOpToArgMinMax::ConvertReduceOpToArgMinMax;

  bool IsValueInitValue(const DenseElementsAttr& attr) const override {
    auto element_type = attr.getType().getElementType();
    if (attr.getNumElements() != 1 || !element_type.isIntOrFloat())
      return false;
    if (mlir::isa<FloatType>(element_type)) {
      auto value = *attr.value_begin<APFloat>();