// and set the device name. It's effectively `TFE_OpSetDevice`, but it is faster
// than separately calling it because if the existing op has the same
// `raw_device_name`, it skips parsing and just leave as it is.
TF_CAPI_EXPORT extern void TFE_OpReset(TFE_Op* op_to_reset,
                                       const char* op_or_function_name,
                                       const char* raw_device_name,

}

TEST_P(CppGradients, TestMatMulGrad) {
  // TODO(vnvo2409): Figure out why `gradient_checker` does not work very
  // well with `MatMul` and remove `TestMatMul*` in
  // `mnist_gradients_test` when done.
  GTEST_SKIP();

  float A_vals[] = {1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f, 7.0f, 8.0f, 9.0f};
  int64_t A_dims[] = {3, 3};
  AbstractTensorHandlePtr A;
  {
    AbstractTensorHandle* A_raw;

    op_stack.pop_back();
    auto op_it = state.op_tape.find(op);
    if (op_it == state.op_tape.end()) {
      // It is possible for ops to end up on the stack if they are unrelated to
      // the target; we should just skip them.
      continue;
    }
    auto trace = std::move(op_it->second);
    state.op_tape.erase(op_it);
    std::vector<Gradient*> out_gradients;
    out_gradients.reserve(trace.output_tensor_info.size());

  std::unique_ptr<TFE_Context, decltype(&TFE_DeleteContext)> context(
      TFE_NewContext(opts.get(), status.get()), TFE_DeleteContext);
  ASSERT_EQ(TF_GetCode(status.get()), TF_OK) << TF_Message(status.get());

  // Skip the test if no TPU is available.
  std::unique_ptr<TF_DeviceList, decltype(&TF_DeleteDeviceList)> devices(
      TFE_ContextListDevices(context.get(), status.get()), TF_DeleteDeviceList);