TF_AbstractTensor* TF_OutputListGet(TF_OutputList* o, int i);
// Append a tensor at the end of the output list, growing its size by one.
void TF_OutputListPushBack(TF_OutputList* o, TF_AbstractTensor* tensor,
                           TF_Status*);

// TF_ExecuteOperation will, if in eager mode, execute, if in graph mode, maybe
// capture some inputs and then add a node in the graph. The output tensors are
// returned through the provided TF_OutputList.

    // Create C++ Tensor
    Tensor src(tensorflow::DT_STRING, TensorShape(dims));
    for (int64_t i = 0; i < src.NumElements(); ++i) {
      src.flat<tstring>()(i) = data[i];
    }
    TF_Tensor* dst = TF_TensorFromTensor(src, &status);
    ASSERT_TRUE(status.ok()) << status.message();

    // Convert back to a C++ Tensor and ensure we get expected output.
    Tensor output;

  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

  TF_Tensor* tensor =
      TF_AllocateTensor(TF_FLOAT, dims.data(), dims.size(), sizeof(float));
  float tensor_data[] = {1};
  memcpy(TF_TensorData(tensor), tensor_data, TF_TensorByteSize(tensor));
  TFE_TensorHandle* tensor_handle = TFE_NewTensorHandle(tensor, status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

  if (reader_ != nullptr) return reader_->DebugString();
  return v2_reader_->DebugString();
}

void CheckpointReader::GetTensor(
    const string& name, std::unique_ptr<tensorflow::Tensor>* out_tensor,
    TF_Status* out_status) const {
  absl::Status status;
  if (reader_ != nullptr) {
    status = reader_->GetTensor(name, out_tensor);
  } else {
    tensorflow::DataType dtype;

                          const char* name) {
  unique_tensor_ptr tensor(BoolTensor(v), TF_DeleteTensor);
  return Const(tensor.get(), graph, s, name);
}

TF_Operation* ScalarConst(int32_t v, TF_Graph* graph, TF_Status* s,
                          const char* name) {
  unique_tensor_ptr tensor(Int32Tensor(v), TF_DeleteTensor);
  return Const(tensor.get(), graph, s, name);
}

                      ", dtype=", DataType_Name(DataType()),
                      ", type=", FullType().DebugString(), ")");
}

absl::Status AbstractTensorHandle::TensorHandleStatus() const {
  // Tensor handles in current runtime don't carry error info and this method
  // should always return OK status.
  return absl::OkStatus();
}

                                                  TF_Status* s) {
  auto handle = dyn_cast<ImmediateExecutionTensorHandle>(unwrap(at));
  if (!handle) {
    string msg = absl::StrCat("Not an eager tensor handle.",
                              reinterpret_cast<uintptr_t>(at));
    TF_SetStatus(s, TF_INVALID_ARGUMENT, msg.c_str());
    return nullptr;
  }
  return wrap(handle);
}

// that defines two inputs, "concat_dim" and "values" (in that order).
// You must use TF_AddInput() for the first input (since it takes a
// single tensor), and TF_AddInputList() for the second input (since
// it takes a list, even if you were to pass a list with a single
// tensor), as in:
//   TF_OperationDescription* desc = TF_NewOperation(graph, "Concat", "c");
//   TF_Output concat_dim_input = {...};

  }
};

// An abstract operation describes an operation by its type, name, and
// attributes. It can be "executed" by the context with some input tensors.
// It is allowed to reusing the same abstract operation for multiple execution
// on a given context, with the same or different input tensors.
class TracingOperation : public AbstractOperation {
 protected:
  explicit TracingOperation(AbstractOperationKind kind)

    *   Transpose now supports 6D tensors.
    *   Float LSTM now supports diagonal recurrent tensors: https://arxiv.org/abs/1903.08023

*   `tf.experimental.dtensor`:

    *   Coordination service now works with `dtensor.initialize_accelerator_system`, and enabled by default.
    *   Add `tf.experimental.dtensor.is_dtensor` to check if a tensor is a DTensor instance.

*   `tf.data`: