def test_accumulator_detached():
mpd = MeanPairwiseDistance()
y_pred = torch.tensor([[3.0, 4.0], [-3.0, -4.0]], requires_grad=True)
y = torch.zeros(2, 2)
mpd.update((y_pred, y))
assert not mpd._sum_of_distances.requires_grad
def test_compute():
mpd = MeanPairwiseDistance()
y_pred = torch.Tensor([[3.0, 4.0], [-3.0, -4.0]])
y = torch.zeros(2, 2)
mpd.update((y_pred, y))
assert mpd.compute() == approx(5.0)
mpd.reset()
y_pred = torch.Tensor([[4.0, 4.0, 4.0, 4.0], [-4.0, -4.0, -4.0, -4.0]])
y = torch.zeros(2, 4)
mpd.update((y_pred, y))
assert mpd.compute() == approx(8.0)
def _test_distrib_accumulator_device(device):
metric_devices = [torch.device("cpu")]
if device.type != "xla":
metric_devices.append(idist.device())
for metric_device in metric_devices:
mpd = MeanPairwiseDistance(device=metric_device)
for dev in [mpd._device, mpd._sum_of_distances.device]:
assert dev == metric_device, f"{type(dev)}:{dev} vs {type(metric_device)}:{metric_device}"
y_pred = torch.Tensor([[3.0, 4.0], [-3.0, -4.0]])
y = torch.zeros(2, 2)
mpd.update((y_pred, y))
for dev in [mpd._device, mpd._sum_of_distances.device]:
assert dev == metric_device, f"{type(dev)}:{dev} vs {type(metric_device)}:{metric_device}"
