Пример #1
0
def test_wrong_input_shapes():
    m = GeometricMeanAbsoluteError()

    with pytest.raises(ValueError):
        m.update((torch.rand(4, 1, 2), torch.rand(4, 1)))

    with pytest.raises(ValueError):
        m.update((torch.rand(4, 1), torch.rand(4, 1, 2)))

    with pytest.raises(ValueError):
        m.update(
            (
                torch.rand(4, 1, 2),
                torch.rand(
                    4,
                ),
            )
        )

    with pytest.raises(ValueError):
        m.update(
            (
                torch.rand(
                    4,
                ),
                torch.rand(4, 1, 2),
            )
        )
Пример #2
0
def test_wrong_input_shapes():
    m = GeometricMeanAbsoluteError()

    with pytest.raises(
            ValueError,
            match=r"Input data shapes should be the same, but given"):
        m.update((torch.rand(4, 1, 2), torch.rand(4, 1)))

    with pytest.raises(
            ValueError,
            match=r"Input data shapes should be the same, but given"):
        m.update((torch.rand(4, 1), torch.rand(4, 1, 2)))

    with pytest.raises(
            ValueError,
            match=r"Input data shapes should be the same, but given"):
        m.update((
            torch.rand(4, 1, 2),
            torch.rand(4, ),
        ))

    with pytest.raises(
            ValueError,
            match=r"Input data shapes should be the same, but given"):
        m.update((
            torch.rand(4, ),
            torch.rand(4, 1, 2),
        ))
def test_zero_sample():
    m = GeometricMeanAbsoluteError()
    with pytest.raises(
            NotComputableError,
            match=
            r"GeometricMeanAbsoluteError must have at least one example before it can be computed"
    ):
        m.compute()
    def _test(n_epochs, metric_device):
        metric_device = torch.device(metric_device)
        n_iters = 80
        s = 16
        n_classes = 2

        offset = n_iters * s
        y_true = torch.rand(size=(offset *
                                  idist.get_world_size(), )).to(device)
        y_preds = torch.rand(size=(offset *
                                   idist.get_world_size(), )).to(device)

        def update(engine, i):
            return (
                y_preds[i * s + rank * offset:(i + 1) * s + rank * offset],
                y_true[i * s + rank * offset:(i + 1) * s + rank * offset],
            )

        engine = Engine(update)

        m = GeometricMeanAbsoluteError(device=metric_device)
        m.attach(engine, "gmae")

        data = list(range(n_iters))
        engine.run(data=data, max_epochs=n_epochs)

        assert "gmae" in engine.state.metrics

        res = engine.state.metrics["gmae"]

        np_y_true = y_true.cpu().numpy()
        np_y_preds = y_preds.cpu().numpy()

        sum_errors = (np.log(np.abs(np_y_true - np_y_preds))).sum()
        np_len = len(y_preds)
        np_ans = np.exp(sum_errors / np_len)

        assert pytest.approx(res) == np_ans
def test_compute():
    a = np.random.randn(4)
    b = np.random.randn(4)
    c = np.random.randn(4)
    d = np.random.randn(4)
    ground_truth = np.random.randn(4)
    np_prod = 1.0

    m = GeometricMeanAbsoluteError()
    m.update((torch.from_numpy(a), torch.from_numpy(ground_truth)))

    errors = np.abs(ground_truth - a)
    np_prod = np.multiply.reduce(errors) * np_prod
    np_len = len(a)
    np_ans = np.power(np_prod, 1.0 / np_len)
    assert m.compute() == pytest.approx(np_ans)

    m.update((torch.from_numpy(b), torch.from_numpy(ground_truth)))
    errors = np.abs(ground_truth - b)
    np_prod = np.multiply.reduce(errors) * np_prod
    np_len += len(b)
    np_ans = np.power(np_prod, 1.0 / np_len)
    assert m.compute() == pytest.approx(np_ans)

    m.update((torch.from_numpy(c), torch.from_numpy(ground_truth)))
    errors = np.abs(ground_truth - c)
    np_prod = np.multiply.reduce(errors) * np_prod
    np_len += len(c)
    np_ans = np.power(np_prod, 1.0 / np_len)
    assert m.compute() == pytest.approx(np_ans)

    m.update((torch.from_numpy(d), torch.from_numpy(ground_truth)))
    errors = np.abs(ground_truth - d)
    np_prod = np.multiply.reduce(errors) * np_prod
    np_len += len(d)
    np_ans = np.power(np_prod, 1.0 / np_len)
    assert m.compute() == pytest.approx(np_ans)
    def _test(y_pred, y, batch_size):
        def update_fn(engine, batch):
            idx = (engine.state.iteration - 1) * batch_size
            y_true_batch = np_y[idx:idx + batch_size]
            y_pred_batch = np_y_pred[idx:idx + batch_size]
            return torch.from_numpy(y_pred_batch), torch.from_numpy(
                y_true_batch)

        engine = Engine(update_fn)

        m = GeometricMeanAbsoluteError()
        m.attach(engine, "gmae")

        np_y = y.numpy()
        np_y_pred = y_pred.numpy()

        data = list(range(y_pred.shape[0] // batch_size))
        gmae = engine.run(data, max_epochs=1).metrics["gmae"]

        sum_errors = (np.log(np.abs(np_y - np_y_pred))).sum()
        np_len = len(y_pred)
        np_ans = np.exp(sum_errors / np_len)

        assert np_ans == pytest.approx(gmae)
    def _test(metric_device):
        metric_device = torch.device(metric_device)
        m = GeometricMeanAbsoluteError(device=metric_device)
        torch.manual_seed(10 + rank)

        y_pred = torch.randint(0, 10, size=(10, ), device=device).float()
        y = torch.randint(0, 10, size=(10, ), device=device).float()

        m.update((y_pred, y))

        # gather y_pred, y
        y_pred = idist.all_gather(y_pred)
        y = idist.all_gather(y)

        np_y_pred = y_pred.cpu().numpy()
        np_y = y.cpu().numpy()

        res = m.compute()

        sum_errors = (np.log(np.abs(np_y - np_y_pred))).sum()
        np_len = len(y_pred)
        np_ans = np.exp(sum_errors / np_len)

        assert np_ans == pytest.approx(res)
def test_zero_div():
    m = GeometricMeanAbsoluteError()
    with pytest.raises(NotComputableError):
        m.compute()