list->add_s(nested_value.s());
          break;
        case AttrValue::kType:
          list->add_type(nested_value.type());
          break;
        case AttrValue::kShape:
          *list->add_shape() = nested_value.shape();
          break;
        default:
          return errors::Unimplemented("Unhandled nested attribute!");
      }
    } else if (auto attr = mlir::dyn_cast<mlir::ElementsAttr>(a)) {

                       AttrValueMap* values) {
  AttrValue value;
  auto& shape_list = *value.mutable_list();
  for (const std::optional<llvm::ArrayRef<int64_t>>& shape : shapes) {
    TensorShapeProto& tshape = *shape_list.add_shape();
    if (shape.has_value()) {
      for (int64_t dim : *shape) {
        tshape.add_dim()->set_size(mlir::ShapedType::isDynamic(dim) ? -1 : dim);
      }
    } else {
      tshape.set_unknown_rank(true);
    }

        handle_dtypes_attr.mutable_list()->add_type(dtype);

        SetTensorShapeProto(subtype,
                            handle_shapes_attr.mutable_list()->add_shape());
      }

      (*node_def->mutable_attr())["_handle_dtypes"] = handle_dtypes_attr;
      (*node_def->mutable_attr())["_handle_shapes"] = handle_shapes_attr;
    }
  }

  TF_RETURN_IF_ERROR(

        handle_dtypes_attr.mutable_list()->add_type(dtype);

        SetTensorShapeProto(subtype,
                            handle_shapes_attr.mutable_list()->add_shape());
      }

      (*node_def->mutable_attr())["_handle_dtypes"] = handle_dtypes_attr;
      (*node_def->mutable_attr())["_handle_shapes"] = handle_shapes_attr;
    }
  }

  TF_RETURN_IF_ERROR(

        const InferredShape& arg_shape = it->second;
        shape_inference::InferenceContext* context =
            shape_refiner->GetContext(n);

        if (arg_shape.handle_type != DT_INVALID) {
          shape_inference::ShapeHandle handle;
          TF_RETURN_IF_ERROR(context->MakeShapeFromPartialTensorShape(
              arg_shape.handle_shape, &handle));

absl::StatusOr<TensorShape> GetTensorShapeFromXlaArgument(
    const XlaArgument& arg) {
  if (absl::holds_alternative<xla::Shape>(arg.shape)) {
    TensorShape arg_shape;
    TF_RETURN_IF_ERROR(
        XLAShapeToTensorShape(std::get<xla::Shape>(arg.shape), &arg_shape));
    return arg_shape;
  } else {
    return std::get<TensorShape>(arg.shape);
  }
}

Status MaybeRewriteLayoutWithShardedShape(
    mlir::StringAttr sharding,

  auto paddings_value = builder->create<TF::ConstOp>(filter.getLoc(), paddings);
  std::vector<int64_t> pad_shape = {filter_shape[0] + pad_h,
                                    filter_shape[1] + pad_w, filter_shape[2],
                                    filter_shape[3]};
  SmallVector<int64_t, 4> expand_shape(pad_shape.begin(), pad_shape.end());

  auto expand_result_type =

     * @return the builder instance
     * @throws NullPointerException if {@code replacement} is null
     */
    @CanIgnoreReturnValue
    public Builder addEscape(char c, String replacement) {
      checkNotNull(replacement);
      // This can replace an existing character (the builder is re-usable).
      replacementMap.put(c, replacement);
      return this;
    }

     * @return the builder instance
     * @throws NullPointerException if {@code replacement} is null
     */
    @CanIgnoreReturnValue
    public Builder addEscape(char c, String replacement) {
      checkNotNull(replacement);
      // This can replace an existing character (the builder is re-usable).
      replacementMap.put(c, replacement);
      return this;
    }

  %tl = "tf.TensorListFromTensor"(%arg0, %elem_shape) : (tensor<10x8x9xf32>, tensor<2xi32>) -> tensor<!tf_type.variant<tensor<8x9xf32>>>
  // CHECK: %[[IND_SHAPE:.*]] = "tf.Const"() <{value = dense<[5, 1]> : tensor<2xi32>}> : () -> tensor<2xi32>
  // CHECK: %[[IND_RESHPE:.*]] = "tf.Reshape"(%[[ARG1]], %[[IND_SHAPE]]) : (tensor<5xi32>, tensor<2xi32>) -> tensor<5x1xi32>