const float* values, int num_values) {
  auto s =
      tensorflow::unwrap(op)->SetAttrFloatList(attr_name, values, num_values);
  if (!s.ok()) {
    LOG(WARNING) << "Unable to set attribute: " << attr_name;
  }
}

void TFE_OpSetAttrIntList(TFE_Op* op, const char* attr_name,
                          const int64_t* values, int num_values) {
  auto s =

  TF_DeleteTensor(t);

  for (int i = 0; i < expected_values.size(); i++) {
    EXPECT_EQ(expected_values[i], actual_values[i])
        << "Mismatch in expected values at (zero-based) index " << i;
  }
}

void CheckRemoteMatMulExecutesOK(TFE_Context* ctx,
                                 const char* remote_device_name,
                                 const char* local_device_name) {

  TF_Tensor* ret;
  TF_SessionRun(session, /*run_options*/ nullptr,
                // input related parameters
                /*inputs*/ nullptr, /*input_values*/ nullptr, /*ninputs*/ 0,
                // output related parameters
                /*outputs*/ &output, /*output_values*/ &ret,
                /*noutputs*/ 1,
                /*targets*/ nullptr, /*ntargets*/ 0,
                /*run_metadata*/ nullptr, status);

  EXPECT_EQ(2, TF_NumDims(out));
  EXPECT_EQ(4, TF_Dim(out, 0));
  EXPECT_EQ(1, TF_Dim(out, 1));
  float* values = static_cast<float*>(TF_TensorData(out));
  // These values are defined to be (input / 2) + 2.
  EXPECT_EQ(2, values[0]);
  EXPECT_EQ(2.5, values[1]);
  EXPECT_EQ(3, values[2]);
  EXPECT_EQ(3.5, values[3]);

  csession.CloseAndDelete(s);
  EXPECT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

    ::testing::Combine(::testing::Values("graphdef", "mlir"),
                       /*tfrt*/ ::testing::Values(false),
                       /*use_function*/ ::testing::Values(true, false)));
#else
INSTANTIATE_TEST_SUITE_P(
    UnifiedCAPI, CppGradients,
    ::testing::Combine(::testing::Values("graphdef", "mlir"),
                       /*tfrt*/ ::testing::Values(false),

  // Returns a Boolean hint indicating whether callers should prefer
  // `SummarizeValue` to resolving this handle and formatting the tensor.
  //
  // For example some tensor handles may represent distributed values, in which
  // case placement information is lost when resolving the handle.
  //
  // If false, a caller might implement pretty-printing by resolving and

  *tensor =
      unwrap(TF_CreateAbstractTensorFromEagerTensor(input_eager, status.get()));
  return absl::OkStatus();
}

// Return a scalar tensor handle with given value.
template <class T, TF_DataType datatype>
Status TestScalarTensorHandle(AbstractContext* ctx, const T value,
                              AbstractTensorHandle** tensor) {
  std::unique_ptr<TF_Status, decltype(&TF_DeleteStatus)> status(
      TF_NewStatus(), TF_DeleteStatus);

//
// The caller retains ownership of `input_values` (which can be deleted using
// TF_DeleteTensor). The caller also retains ownership of `run_options` and/or
// `run_metadata` (when not NULL) and should manually call TF_DeleteBuffer on
// them.
//
// On success, the tensors corresponding to outputs[0,noutputs-1] are placed in
// output_values[]. Ownership of the elements of output_values[] is transferred

// - graphdef and MLIR tracing engine
INSTANTIATE_TEST_SUITE_P(Tracing, UnifiedCAPI,
                         ::testing::Combine(::testing::Values("graphdef",
                                                              "mlir"),
                                            ::testing::Values(false)));

}  // namespace

  std::vector<TF_Operation*> return_nodes;
  std::vector<const char*> missing_unused_key_names;
  std::vector<int> missing_unused_key_indexes;

  // Backing memory for missing_unused_key_names values.
  std::vector<tensorflow::string> missing_unused_key_names_data;
};

struct TF_DeviceList {
  std::vector<tensorflow::DeviceAttributes> response;
};

struct TF_Function {