Status BuildImmediateExecutionContext(bool use_tfrt, AbstractContext** ctx);

// Return a tensor handle with given type, values and dimensions.
template <class T, TF_DataType datatype>
Status TestTensorHandleWithDims(AbstractContext* ctx, const T* data,
                                const int64_t* dims, int num_dims,
                                AbstractTensorHandle** tensor) {

                    absl::Span<AbstractTensorHandle*> outputs) {
  return ops::BiasAdd(ctx, inputs[0], inputs[1], &outputs[0], "NHWC",
                      "BiasAdd");
}

class CppGradients
    : public ::testing::TestWithParam<std::tuple<const char*, bool, bool>> {
 protected:
  void SetUp() override {
    TF_StatusPtr status(TF_NewStatus());

  ptrs->reset(new const void*[v.size()]);
  lens->reset(new size_t[v.size()]);
  for (size_t i = 0; i < v.size(); ++i) {
    (*ptrs)[i] = v[i].data();
    (*lens)[i] = v[i].size();
  }
}

class CApiColocationTest : public ::testing::Test {
 protected:
  CApiColocationTest() : s_(TF_NewStatus()), graph_(TF_NewGraph()) {}

  void SetUp() override {
    feed1_ = Placeholder(graph_, s_, "feed1");

#include <cstdint>
#include <string>

#include "tensorflow/core/platform/status.h"

namespace tsl {
class CoordinationServiceAgent;
}

namespace tensorflow {
class ImmediateExecutionContext;
class ServerDef;
class WorkerEnv;
class WorkerCacheInterface;

class ImmediateExecutionDistributedManager {
 public:
  virtual ~ImmediateExecutionDistributedManager() {}

using std::vector;
using tensorflow::ops::BiasAddGrad;
using tensorflow::ops::Mul;
using tensorflow::ops::ReluGrad;

namespace tensorflow {
namespace gradients {
namespace {

class ReluGradientFunction : public GradientFunction {
 public:
  explicit ReluGradientFunction(vector<AbstractTensorHandle*> f_outputs)
      : forward_outputs_(f_outputs) {
    for (auto output : forward_outputs_) {

#include "tensorflow/core/platform/refcount.h"
namespace tensorflow {
namespace tracing {
namespace graph {
using tensorflow::AbstractFunction;
// Thin wrapper around a FunctionDef.
class GraphFunction : public AbstractFunction {
 public:
  explicit GraphFunction(FunctionDef fdef);
  ~GraphFunction() override;

  // GraphFunction maybe stay alive for the duration of the returned
  // FunctionDef.

  // for computing gradient so we could safely drop it.
  outputs[0] = temp_outputs[1];
  temp_outputs[0]->Unref();
  return absl::OkStatus();
}

class CppGradients
    : public ::testing::TestWithParam<std::tuple<const char*, bool, bool>> {
 protected:
  void SetUp() override {
    TF_StatusPtr status(TF_NewStatus());

    for details.

#### Running doctest for testable docstring

There are two ways to test the code in the docstring locally:

1.  If you are only changing the docstring of a class/function/method, then you
    can test it by passing that file's path to
    [tf_doctest.py](https://www.tensorflow.org/code/tensorflow/tools/docs/tf_doctest.py).
    For example:

    ```bash

tensorflow::DataType TapeTensor::GetDType() const {
  return handle_->DataType();
}
AbstractTensorHandle* TapeTensor::GetHandle() const { return handle_; }

AbstractTensorHandle* TapeTensor::ZerosLike() const { return nullptr; }

class TapeVSpace
    : public eager::VSpace<AbstractTensorHandle, GradientFunction, TapeTensor> {
 public:
  explicit TapeVSpace(AbstractContext* ctx) : ctx_(ctx) {}
  ~TapeVSpace() override {}

             BaseType(dtype) != DT_VARIANT) {
    return errors::InvalidArgument(
        "Expected numeric or variant tensor, got dtype ", dtype);
  }
  return tensorflow::ops::Conj(ctx, input, output, name);
}

class AddGradientFunction : public GradientFunction {
 public:
  Status Compute(AbstractContext* ctx,
                 absl::Span<AbstractTensorHandle* const> grad_outputs,