build(deps): bump mypy from 1.11.2 to 1.13.0 in /requirements (#2808)

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Jirka B <[email protected]> Co-authored-by: Jirka Borovec <[email protected]> (cherry picked from commit 421e028)
Lightning-AI · Nov 7, 2024 · 3b6b9a2 · 3b6b9a2
1 parent 7f0fafc
commit 3b6b9a2
Show file tree

Hide file tree

Showing 4 changed files with 11 additions and 10 deletions.
diff --git a/requirements/typing.txt b/requirements/typing.txt
@@ -1,4 +1,4 @@
-mypy ==1.11.2
+mypy ==1.13.0
 torch ==2.5.0
 
 types-PyYAML

diff --git a/src/torchmetrics/detection/_mean_ap.py b/src/torchmetrics/detection/_mean_ap.py
@@ -849,9 +849,9 @@ def __calculate_recall_precision_scores(
 
             inds = torch.searchsorted(rc, rec_thresholds.to(rc.device), right=False)
             num_inds = inds.argmax() if inds.max() >= tp_len else num_rec_thrs
-            inds = inds[:num_inds]  # type: ignore[misc]
-            prec[:num_inds] = pr[inds]  # type: ignore[misc]
-            score[:num_inds] = det_scores_sorted[inds]  # type: ignore[misc]
+            inds = inds[:num_inds]
+            prec[:num_inds] = pr[inds]
+            score[:num_inds] = det_scores_sorted[inds]
             precision[idx, :, idx_cls, idx_bbox_area, idx_max_det_thresholds] = prec
             scores[idx, :, idx_cls, idx_bbox_area, idx_max_det_thresholds] = score
 

diff --git a/src/torchmetrics/functional/audio/pit.py b/src/torchmetrics/functional/audio/pit.py
@@ -182,7 +182,7 @@ def permutation_invariant_training(
         metric_of_ps = metric_func(ppreds, ptarget)
         metric_of_ps = torch.mean(metric_of_ps.reshape(batch_size, len(perms), -1), dim=-1)
         # find the best metric and best permutation
-        best_metric, best_indexes = eval_op(metric_of_ps, dim=1)  # type: ignore[call-overload]
+        best_metric, best_indexes = eval_op(metric_of_ps, dim=1)
         best_indexes = best_indexes.detach()
         best_perm = perms[best_indexes, :]
         return best_metric, best_perm

diff --git a/src/torchmetrics/wrappers/tracker.py b/src/torchmetrics/wrappers/tracker.py
@@ -219,7 +219,8 @@ def best_metric(
     ) -> Union[
         None,
         float,
-        Tuple[float, int],
+        Tensor,
+        Tuple[Union[int, float, Tensor], Union[int, float, Tensor]],
         Tuple[None, None],
         Dict[str, Union[float, None]],
         Tuple[Dict[str, Union[float, None]], Dict[str, Union[int, None]]],
@@ -260,7 +261,7 @@ def best_metric(
         if isinstance(self._base_metric, Metric):
             fn = torch.max if self.maximize else torch.min
             try:
-                value, idx = fn(res, 0)  # type: ignore[call-overload]
+                value, idx = fn(res, 0)
                 if return_step:
                     return value.item(), idx.item()
                 return value.item()
@@ -277,11 +278,11 @@ def best_metric(
 
         else:  # this is a metric collection
             maximize = self.maximize if isinstance(self.maximize, list) else len(res) * [self.maximize]
-            value, idx = {}, {}
+            value, idx = {}, {}  # type: ignore[assignment]
             for i, (k, v) in enumerate(res.items()):
                 try:
                     fn = torch.max if maximize[i] else torch.min
-                    out = fn(v, 0)  # type: ignore[call-overload]
+                    out = fn(v, 0)
                     value[k], idx[k] = out[0].item(), out[1].item()
                 except (ValueError, RuntimeError) as error:  # noqa: PERF203 # todo
                     rank_zero_warn(
@@ -290,7 +291,7 @@ def best_metric(
                         "Returning `None` instead.",
                         UserWarning,
                     )
-                    value[k], idx[k] = None, None
+                    value[k], idx[k] = None, None  # type: ignore[assignment]
 
             if return_step:
                 return value, idx