Move checking in r2 for number of samples to compute (#426)

* fix

* changelog

CHANGELOG.md

@@ -56,6 +56,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Use `torch.argmax` instead of `torch.topk` when `k=1` for better performance ([#419](https://github.com/PyTorchLightning/metrics/pull/419))
 
 
+- Moved check for number of samples in R2 score to support single sample updating ([#426](https://github.com/PyTorchLightning/metrics/pull/426))
+
+
 ### Deprecated
 
 - Rename `r2score` >> `r2_score` and `kldivergence` >> `kl_divergence` in `functional` ([#371](https://github.com/PyTorchLightning/metrics/pull/371))

tests/regression/test_r2.py

@@ -143,6 +143,12 @@ def test_error_on_too_few_samples(metric_class=R2Score):
     metric = metric_class()
     with pytest.raises(ValueError, match="Needs at least two samples to calculate r2 score."):
         metric(torch.randn(1), torch.randn(1))
+    metric.reset()
+
+    # calling update twice should still work
+    metric.update(torch.randn(1), torch.randn(1))
+    metric.update(torch.randn(1), torch.randn(1))
+    assert metric.compute()
 
 
 def test_warning_on_too_large_adjusted(metric_class=R2Score):

torchmetrics/functional/regression/r2.py

@@ -34,8 +34,6 @@ def _r2_score_update(preds: Tensor, target: Tensor) -> Tuple[Tensor, Tensor, Ten
             "Expected both prediction and target to be 1D or 2D tensors,"
             f" but received tensors with dimension {preds.shape}"
         )
-    if len(preds) < 2:
-        raise ValueError("Needs at least two samples to calculate r2 score.")
 
     sum_obs = torch.sum(target, dim=0)
     sum_squared_obs = torch.sum(target * target, dim=0)
@@ -77,6 +75,9 @@ def _r2_score_compute(
         >>> _r2_score_compute(sum_squared_obs, sum_obs, rss, n_obs, multioutput="raw_values")
         tensor([0.9654, 0.9082])
     """
+    if n_obs < 2:
+        raise ValueError("Needs at least two samples to calculate r2 score.")
+
     mean_obs = sum_obs / n_obs
     tss = sum_squared_obs - sum_obs * mean_obs
     raw_scores = 1 - (rss / tss)