}

namespace {

// Helper method that creates a shape handle for a shape described by dims.
tensorflow::shape_inference::ShapeHandle ShapeHandleFromDims(
    tensorflow::shape_inference::InferenceContext* ic, int num_dims,
    const int64_t* dims) {
  if (num_dims != -1) {
    std::vector<tensorflow::shape_inference::DimensionHandle> dim_vec;
    dim_vec.reserve(num_dims);
    for (int i = 0; i < num_dims; ++i) {

struct TF_Library {
  void* lib_handle;
  TF_Buffer op_list;
};

struct TF_Graph {
  TF_Graph();

  mutable tensorflow::mutex mu;
  tensorflow::Graph graph TF_GUARDED_BY(mu);

  // Runs shape inference.
  tensorflow::ShapeRefiner refiner TF_GUARDED_BY(mu);

  // Maps from name of an operation to the Node* in 'graph'.
  std::unordered_map<tensorflow::string, tensorflow::Node*> name_map
      TF_GUARDED_BY(mu);

  using tensorflow::NodeDef;
  using tensorflow::OpRegistrationData;
  using tensorflow::Tensor;
  using tensorflow::shape_inference::DimensionHandle;
  using tensorflow::shape_inference::InferenceContext;
  using tensorflow::shape_inference::ShapeAndType;
  using tensorflow::shape_inference::ShapeHandle;

  const int num_inputs = input_shapes->num_items;
  NodeDef node_def;

// `shape_inference::InferenceContext` constructor. Note the following:
//   - The inputs of the `op` are not used for shape inference. So, it is
//     OK to not have the inputs properly set in `op`. See `input_tensors`
//     if you want shape inference to consider the input tensors of the
//     op for shape inference.
//   - The types need not be set in `input_shapes` as it is not used.

      .Attr("v: " #type)                                      \
      .SetShapeFn(tensorflow::shape_inference::UnknownShape); \
  REGISTER_OP("CApiAttributesTestOpList" #type)               \
      .Attr("v: list(" #type ")")                             \
      .SetShapeFn(tensorflow::shape_inference::UnknownShape)
ATTR_TEST_REGISTER_OP(string);
ATTR_TEST_REGISTER_OP(int);
ATTR_TEST_REGISTER_OP(float);

### 2. Code

Provide code to help us reproduce your issues using one of the following options:

#### Option A: Reference colab notebooks

1)  Reference [TensorFlow Model Colab](https://colab.research.google.com/gist/ymodak/e96a4270b953201d5362c61c1e8b78aa/tensorflow-datasets.ipynb?authuser=1): Demonstrate how to build your TF model.

“parameter server” designs the management of shared state is built into the system, TensorFlow enables developers to experiment with novel optimizations and training algorithms. TensorFlow supports a variety of applications, with a focus on training and inference on deep neural networks. Several Google services use TensorFlow in production, we have released it as an open-source project, and it has become widely used for machine learning research. In this paper, we describe the TensorFlow dataflow model and...

  F release() {
    released_ = true;
    return std::move(f_);
  }

  bool is_released() const { return released_; }

 private:
  static_assert(!std::is_reference<F>::value, "F must not be a reference");

  bool released_ = false;
  F f_;
};

template <int&... ExplicitParameterBarrier, typename F,
          typename DecayF = typename std::decay<F>::type>

  ~GraphFunction() override;

  // GraphFunction maybe stay alive for the duration of the returned
  // FunctionDef.
  Status GetFunctionDef(const FunctionDef** fdef) override;

  // Returns a shared reference to the wrapped function.
  absl::StatusOr<core::RefCountPtr<FunctionRecord>> GetFunctionRecord()
      override {
    return func_record_.GetNewRef();
  }

  // For LLVM style RTTI.

TF_CAPI_EXPORT extern int TFE_OpGetFlatInputCount(const TFE_Op* op,
                                                  TF_Status* status);
// Returns a borrowed reference to one of `op`'s inputs. Use
// `TFE_TensorHandleCopySharingTensor` to make a new reference.
TF_CAPI_EXPORT extern TFE_TensorHandle* TFE_OpGetFlatInput(const TFE_Op* op,
                                                           int index,