if (TF_GetCode(s) != TF_OK) {
      return;
    }
  }
  int num_outputs = unwrap(o)->expected_num_outputs;
  tsl::Set_TF_Status_from_Status(
      s, unwrap(op)->Execute(
             absl::MakeSpan(reinterpret_cast<AbstractTensorHandle**>(
                                unwrap(o)->outputs.data()),
                            unwrap(o)->outputs.size()),
             &num_outputs));
}

  TFE_DeleteTensorHandle(h1_task1);
  TFE_DeleteTensorHandle(retvals[0]);

  TFE_DeleteOp(matmul);

  TFE_Executor* executor = TFE_ContextGetExecutorForThread(ctx);
  TFE_ExecutorWaitForAllPendingNodes(executor, status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  TFE_DeleteExecutor(executor);
  TFE_DeleteContext(ctx);

  TF_DeleteStatus(status);

  custom_device.copy_tensor_from_device = &CopyTensorFromLoggingDevice;
  custom_device.delete_device = &DeleteLoggingDevice;
  custom_device.execute = &LoggingDeviceExecute;
  LoggingDevice* device = new LoggingDevice;
  device->arrived_flag = arrived_flag;
  device->executed_flag = executed_flag;
  device->device_name = name;
  device->underlying_device = "/job:localhost/replica:0/task:0/device:CPU:0";

  GradientRegistry registry;
  s = RegisterGradients(&registry);
  ASSERT_EQ(errors::OK, s.code()) << s.message();
  auto tape = std::make_unique<Tape>(/*persistent=*/false);
  s = Execute(check_numerics_op.get(), ctx.get(), absl::MakeSpan(outputs),
              &num_retvals, &forward_op, tape.get(), registry);
  ASSERT_EQ(errors::OK, s.code()) << s.message();

  string read_message;

      TF_RETURN_IF_ERROR(fn_op->AddInput(input));
    }
    int retvals = outputs.size() - null_indices.size();
    std::vector<AbstractTensorHandle*> fn_outputs(retvals);
    TF_RETURN_IF_ERROR(fn_op->Execute(
        absl::Span<AbstractTensorHandle*>(fn_outputs.data(), fn_outputs.size()),
        &retvals));
    int skipped_indices = 0;
    for (int i = 0; i < outputs.size(); i++) {
      if (!null_indices.contains(i)) {

  for (auto* h : handles_task2) {
    TFE_DeleteTensorHandle(h);
  }

  TFE_DeleteOp(matmul);

  TFE_Executor* executor = TFE_ContextGetExecutorForThread(ctx);
  TFE_ExecutorWaitForAllPendingNodes(executor, status);
  ASSERT_EQ(TF_GetCode(status), TF_OK) << TF_Message(status);
  TFE_DeleteExecutor(executor);
  if (func) {
    TFE_ContextRemoveFunction(ctx, "MatMulFunction", status);
  }
  TFE_DeleteContext(ctx);