*b_map = {};
  *b_size = {};
  int block_rank = 0;
  for (int i = 0; i < dims_count; i++) {
    if (block_size[i] != 1) {
      traversal_order->push_back(block_rank + dims_count);
      format->push_back(kTfLiteDimDense);
      block_rank++;
      b_map->push_back(i);
      b_size->push_back(block_size[i]);
    }
  }
}

inline float GetSparsity(const int num_zeros, const int num_elements) {

  args.reserve(real_argc + 1);
  for (int i = 0; i < real_argc; i++) {
    args.push_back(real_argv[i]);
  }

  struct FreeDeleter {
    void operator()(char* ptr) { free(ptr); }
  };

  std::unique_ptr<char, FreeDeleter> enable_global_jit_arg(
      strdup("--tf_xla_cpu_global_jit=true"));
  args.push_back(enable_global_jit_arg.get());

  std::unique_ptr<char, FreeDeleter> reduce_min_cluster_size_arg(

        auto newTy = QuantizedType::castToExpressedType(operand.getType());
        newOperands.push_back(
            rewriter.create<TFL::DequantizeOp>(loc, newTy, operand));
        continue;
      }

      newOperands.push_back(operand);
    }

    SmallVector<Type> newResultTys;
    for (auto result : op->getResults()) {
      Type resultTy = result.getType();

  result.reserve(set.size());
  for (DataType dt : set) {
    result.push_back(dt);
  }
  return result;
}

std::vector<std::vector<int64_t>> InterestingShapes() {
  std::vector<std::vector<int64_t>> interesting_shapes;
  interesting_shapes.push_back({});             // Scalar
  interesting_shapes.push_back({10});           // 1D Vector
  interesting_shapes.push_back({3, 3});         // 2D Matrix

      padded_begin.push_back(0);
      padded_end.push_back(0);
      padded_stride.push_back(1);
    }

    // Account for ellipsis mask.
    ++index;

    // After the ellipsis.
    for (; index < begin_shape[0];) {
      padded_begin.push_back(begin_dense_elem_attr.getValues<int32_t>()[index]);
      padded_end.push_back(end_dense_elem_attr.getValues<int32_t>()[index]);

    for (const int64_t dim : batch_dimensions) {
      batch_dimensions_.axes.push_back(dim);
      batch_dimensions_.sizes.push_back(type.getDimSize(dim));
    }

    for (const int64_t dim : contracting_dimensions) {
      contracting_dimensions_.axes.push_back(dim);
      contracting_dimensions_.sizes.push_back(type.getDimSize(dim));
    }

    for (int64_t dim = 0; dim < rank; ++dim) {

        if (mlir::isa<NoneType>(operand_type)) {
          inputs.push_back(operand);
          continue;
        }

        auto ele_type =
            mlir::cast<TensorType>(operand.getType()).getElementType();
        if (auto dq_op =
                dyn_cast_or_null<DequantizeOpT>(operand.getDefiningOp())) {
          inputs.push_back(dq_op.getOperand());
        } else if (!ele_type.isF32()) {

    });

    std::vector<std::string> lines;
    lines.push_back("-------- Quantization Summary --------");
    lines.push_back("Number of quantized layers in the model");
    lines.push_back("--------------------------------");
    lines.push_back(
        absl::StrFormat("%s Count/Total", pad_string("Name", name_col_width)));
    lines.push_back("================================");

};

TEST_F(PjRtExecutionUtilGpuTest, PreparePjRtExecutableArguments) {
  std::vector<const Tensor*> inputs;
  inputs.push_back(CreateDeviceTensor<int32_t>(TensorShape({1, 3}), {0, 0, 0}));
  inputs.push_back(CreateDeviceTensor<int32_t>(TensorShape({1, 3}), {1, 2, 3}));
  inputs.push_back(CreateDeviceTensor<int32_t>(TensorShape({1, 3}), {4, 5, 6}));
  int num_missing_prefix_ctx_inputs = 2;
  std::vector<int> input_mapping{3, 4};

    if (aggregator_ops_to_ignore.count(op.getId().str())) continue;

    ids.push_back(op.getId());
    calibration_methods.push_back(op.getCalibrationMethod());
    statistics_outputs.push_back(op.getMin());
    statistics_outputs.push_back(op.getMax());
    statistics_outputs.push_back(op.getHistogram());
  }
}

// Inserts a `CalibrationStatisticsSaverOp` to the end of the region.