const tensorflow::AttrValue& default_value,
                          const char* attr_name, TF_Status* status) {
  switch (default_value.value_case()) {
    case tensorflow::AttrValue::kS: {
      const string& v = default_value.s();
      TFE_OpSetAttrString(op, attr_name, v.data(), v.size());
      break;
    }
    case tensorflow::AttrValue::kI:

      }
    }
    case AttrValue::kB:
      return attr_value.b() ? "true" : "false";
    case AttrValue::kType:
      return DataType_Name(attr_value.type());
    case AttrValue::kShape:
      return PrintTensorShape(attr_value.shape());
    case AttrValue::kTensor:
      return PrintTensorProto(attr_value.tensor());
    case AttrValue::kList: {
      string ret = "{";

  def.add_input("input1");
  def.add_input("input2");
  def.add_input("input3");
  def.add_input("input3");
  def.add_input("input3");

  AttrValue v0;
  v0.set_type(DataType::DT_INT32);
  v0.set_i(3);
  (*def.mutable_attr())["NumInput3"] = v0;
  AttrValue v1;
  v1.set_type(DataType::DT_FLOAT);
  (*def.mutable_attr())["SomeDataTypeAttr"] = v1;

typedef struct TFE_Op TFE_Op;

namespace tensorflow {
DEFINE_CONVERSION_FUNCTIONS(tensorflow::AbstractOpAttrs, TFE_OpAttrs);

// Set an AttrValue on the op. Doesn't handle the list types.
void SetOpAttrValueScalar(TFE_Context* ctx, TFE_Op* op,
                          const tensorflow::AttrValue& default_value,
                          const char* attr_name, TF_Status* status);
}  // namespace tensorflow

  const string& base_type() const { return base_type_; }
  const AttrValue& default_value() const { return default_value_; }
  const bool is_list() const { return is_list_; }

 private:
  explicit AttrSpec(const OpDef::AttrDef& attr_def);

  string name_;
  string description_;
  string full_type_;
  string base_type_;
  AttrValue default_value_;
  bool is_list_;
};

}  // namespace generator

impl::NodeMatcherProperties Attr(std::pair<string, AttrValue> attrs);
impl::NodeMatcherProperties Attr(string name);

std::pair<string, AttrValue> AttrLiteralHelper(
    const std::pair<string, bool>& bool_attr);

std::pair<string, AttrValue> AttrLiteralHelper(
    const std::pair<string, absl::Span<const int>>& int_list_attr);

std::pair<string, AttrValue> AttrLiteralHelper(

  EXPECT_EQ(shape_inference_graphs.size(), 0);

  // Check host graph: verify we have key placeholder and sequencer.
  std::unique_ptr<FunctionBody> host_fbody;
  AttrValue device_ordinal_temp_value;
  device_ordinal_temp_value.set_i(0);
  protobuf::Map<string, AttrValue> host_func_attrs;
  host_func_attrs["_device_ordinal"] = device_ordinal_temp_value;
  TF_CHECK_OK(FunctionDefToBodyHelper(

      LIST_CASE(tensor, TF_ATTR_FUNC);
#undef LIST_CASE
      break;

    case tensorflow::AttrValue::kPlaceholder:
      *list_size = -1;
      *total_size = -1;
      break;

    case tensorflow::AttrValue::kFunc:
      *list_size = -1;
      *total_size = -1;
      break;

    case tensorflow::AttrValue::VALUE_NOT_SET:
      status->status = absl::InvalidArgumentError(

                   TensorShape expected_shape, error::Code expected_code) {
  Status status;
  NodeDef def;
  def.set_op("Bitcast");
  def.set_device(DEVICE_CPU);

  AttrValue typeAttr;
  SetAttrValue(input_tensor->dtype(), &typeAttr);

  AttrValue outTypeAttr;
  SetAttrValue(out_type, &outTypeAttr);

  (*def.mutable_attr())["T"] = typeAttr;
  (*def.mutable_attr())["type"] = outTypeAttr;

  def.add_input(

                boolean attrValue = reader.nextBoolean();
                attributes = attributesFactory.concat(attributes, Attribute.of(attrName, Boolean.class), attrValue);
            } else if (reader.peek() == NUMBER) {
                Integer attrValue = reader.nextInt();
                attributes = attributesFactory.concat(attributes, Attribute.of(attrName, Integer.class), attrValue);
            } else {