status_ = StatusFromTF_Status(status.get());
    ASSERT_EQ(errors::OK, status_.code()) << status_.message();

    {
      AbstractContext* ctx_raw = nullptr;
      status_ =
          BuildImmediateExecutionContext(std::get<1>(GetParam()), &ctx_raw);
      ASSERT_EQ(errors::OK, status_.code()) << status_.message();
      immediate_execution_ctx_.reset(ctx_raw);
    }

  TF_OperationDescription* desc = TF_NewOperation(graph, "Const", name);
  TF_SetAttrTensor(desc, "value", t, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);
  TF_SetAttrType(desc, "dtype", TF_TensorType(t));
  *op = TF_FinishOperation(desc, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);
  ASSERT_NE(*op, nullptr);
}

TF_Operation* Const(TF_Tensor* t, TF_Graph* graph, TF_Status* s,

      if (n == nullptr) continue;

      // Adding the gradients to the graph can alter the prefix to prevent
      // name collisions only if this prefix has not been provided explicitly
      // by the user. If it was provided, assert that it remained intact.
      if (prefix != nullptr && !absl::StartsWith(n->name(), prefix_cmp)) {
        status->status = tensorflow::errors::Internal(

  auto iter = registry_.find(op_name);
  if (iter != registry_.end()) {
    const string error_msg = "Gradient already exists for op: " + op_name + ".";
    return errors::AlreadyExists(error_msg);
  }
  registry_.insert({op_name, gradient_function_factory});
  return absl::OkStatus();
}
Status GradientRegistry::Lookup(
    const ForwardOperation& op,
    std::unique_ptr<GradientFunction>* gradient_function) const {

    delete[] output_names_ptr;
    if (expect_failure) {
      ASSERT_EQ(func_, nullptr);
      return;
    }

    ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
    ASSERT_NE(func_, nullptr);
    ASSERT_EQ(std::string(func_name_), std::string(TF_FunctionName(func_)));
    TF_GraphCopyFunction(host_graph_, func_, nullptr, s_);
    ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
  }

    ASSERT_EQ(errors::OK, s.code()) << s.message();
    s = dyn_cast<tracing::TracingContext>(ctx.get())->AddParameter(
        DT_FLOAT, shape, &x_raw);
    ASSERT_EQ(errors::OK, s.code()) << s.message();
    x.reset(x_raw);
  }

  PartialTensorShape shape;
  Status s = x->Shape(&shape);
  ASSERT_EQ(errors::OK, s.code()) << s.message();
  ASSERT_FALSE(shape.unknown_rank());

                          &numerical_grad_raw);
    ASSERT_EQ(errors::OK, s.code()) << s.message();
    numerical_grad.reset(numerical_grad_raw);
  }

  TF_Tensor* numerical_tensor;
  s = GetValue(numerical_grad.get(), &numerical_tensor);
  ASSERT_EQ(errors::OK, s.code()) << s.message();
  auto num_elem_numerical = TF_TensorElementCount(numerical_tensor);
  ASSERT_EQ(num_elem_numerical, num_grad);

  TF_GraphCopyFunction(func_graph_, func_, nullptr, s_);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

  EXPECT_EQ(TF_GraphNumFunctions(func_graph_), 1);
  EXPECT_EQ(TF_GraphGetFunctions(func_graph_, funcs, 1, s_), 1);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

  TF_GraphRemoveFunction(func_graph_, func_name_, s_);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

        if (model_outputs[i]) {
          output_list.outputs.emplace_back(model_outputs[i]);
          output_list.expected_num_outputs += 1;
        } else {
          null_indices.insert(i);
        }
      }
      TF_RETURN_IF_ERROR(dyn_cast<tracing::TracingContext>(func_ctx.get())
                             ->Finalize(&output_list, &func));
      scoped_func.reset(func);

  tensorflow::JobDef* job_def = cluster_def->add_job();
  job_def->set_name(job_name);
  for (int i = 0; i < num_tasks; i++) {
    int port = tensorflow::testing::PickUnusedPortOrDie();
    job_def->mutable_tasks()->insert(
        {i, tensorflow::strings::StrCat("localhost:", port)});
    LOG(INFO) << "Picked test port: " << port << " for job: " << job_name
              << ", task: " << i;
  }
  return server_def;
}