TF_SetTracingImplementation(std::get<0>(GetParam()), status.get());
    Status s = StatusFromTF_Status(status.get());
    CHECK_EQ(errors::OK, s.code()) << s.message();
  }

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

// Checks that inputs[0] is a scalar.
Status TestScalarShape(AbstractContext* ctx,

    // low tolerances
    enable_tensor_float_32_execution(false);
  }

  AbstractContextPtr immediate_execution_ctx_;
  GradientRegistry registry_;
  Status status_;

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

TEST_P(CppGradients, TestReluGrad) {
  status_ = registry_.Register("Relu", ReluRegisterer);

    // low tolerances
    enable_tensor_float_32_execution(false);
  }

  AbstractContextPtr immediate_execution_ctx_;
  GradientRegistry registry_;
  Status status_;

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

TEST_P(CppGradients, TestIdentityNGrad) {

    // unstable. Some forward pass tests also fail with TensorFloat-32 due to
    // low tolerances
    enable_tensor_float_32_execution(false);
  }

  AbstractContextPtr ctx_;

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

TEST_P(GradientCheckerTest, TestMatMul) {
  float A_vals[] = {1.0f, 2.0f, 3.0f, 4.0f};

    // low tolerances
    enable_tensor_float_32_execution(false);
  }

  AbstractContextPtr immediate_execution_ctx_;
  GradientRegistry registry_;
  Status status_;

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

TEST_P(CppGradients, TestAddGrad) {
  AbstractTensorHandlePtr x;
  {