// `tensor_shape`. `tfl_pad_values` should be formatted as `[[low_0, high_0],
  // [low_1, high_1], ..., [low_n, high_n]]`, where `low_x` and `high_x` are the
  // low and high paddings for the x-th dimension, respectively.
  SmallVector<int64_t> InferPaddedTensorShape(
      const ArrayRef<int64_t> tensor_shape,
      const ArrayRef<int32_t> tfl_pad_values) const {

      std::vector<int64_t> tensor_shape =
          mlir::cast<ShapedType>(tensor.getType()).getShape().vec();

      // Calculate offsets based on depth_idx, channel_idx and tensor_shape
      std::vector<int64_t> start_indices(tensor_shape.size(), 0);
      std::vector<int64_t> limit_indices = tensor_shape;
      const std::vector<int64_t> strides(tensor_shape.size(), 1);
      start_indices[channel_idx] = depth_idx;

#include "tensorflow/core/framework/shape_inference.h"
#include "tensorflow/core/framework/tensor.h"
#include "tensorflow/core/framework/tensor.pb.h"  // NOLINT
#include "tensorflow/core/framework/tensor_shape.h"
#include "tensorflow/core/framework/tensor_shape.pb.h"
#include "tensorflow/core/framework/types.h"
#include "tensorflow/core/framework/versions.pb.h"
#include "tensorflow/core/graph/graph.h"
#include "tensorflow/core/graph/node_builder.h"

    for var_name, shape in var_name_and_shapes:
      self.assertIn(var_name, variable_node_defs)
      self.assertEqual(
          shape,
          tensor_shape.TensorShape(
              variable_node_defs[var_name].attr['shape'].shape
          ),
      )

  @parameterized.named_parameters(
      ('use_constant_with_int32_input', dtypes.int32, False, True),

#include "tensorflow/core/framework/graph_to_functiondef.h"
#include "tensorflow/core/framework/node_def_builder.h"
#include "tensorflow/core/framework/node_def_util.h"
#include "tensorflow/core/framework/tensor_shape.pb.h"
#include "tensorflow/core/graph/algorithm.h"
#include "tensorflow/core/lib/core/errors.h"
#include "tensorflow/core/lib/gtl/cleanup.h"
#include "tensorflow/core/platform/macros.h"

    ArrayInfo array_info;
    array_info.imported_dtype = tensor_info.dtype();

    if (tensor_info.has_tensor_shape()) {
      array_info.shape = tensor_info.tensor_shape();
    } else {
      // If there is no tensor shape in the tensor info, conservatively set
      // unknown_rank to true.
      array_info.shape.set_unknown_rank(true);
    }