Fix error raised by latest Torchmetrics (0.11.0) (#576)

* update default torchmetrics

* unpin torchmetrics version

* update torchmetrics requirement

* update torchmetrics version from Oliver suggestion

requirements/pytorch.txt

@@ -1,2 +1,2 @@
 torch>=1.0
-torchmetrics==0.3.2
+torchmetrics>=0.10.0

tests/torch/test_trainer.py

@@ -325,9 +325,9 @@ def test_evaluate_results(torch_yoochoose_next_item_prediction_model):
         (
             tr.BinaryClassificationTask("click", summary_type="mean"),
             [
-                "eval_/click/binary_classification_task/accuracy",
-                "eval_/click/binary_classification_task/precision",
-                "eval_/click/binary_classification_task/recall",
+                "eval_/click/binary_classification_task/binary_accuracy",
+                "eval_/click/binary_classification_task/binary_precision",
+                "eval_/click/binary_classification_task/binary_recall",
             ],
         ),
         (
@@ -446,9 +446,9 @@ def test_trainer_with_multiple_tasks():
         "eval_/next-item/avg_precision_at_20",
         "eval_/next-item/recall_at_10",
         "eval_/next-item/recall_at_20",
-        "eval_/click/binary_classification_task/accuracy",
-        "eval_/click/binary_classification_task/precision",
-        "eval_/click/binary_classification_task/recall",
+        "eval_/click/binary_classification_task/binary_accuracy",
+        "eval_/click/binary_classification_task/binary_precision",
+        "eval_/click/binary_classification_task/binary_recall",
         "eval_/play_percentage/regression_task/mean_squared_error",
     ]
 

transformers4rec/torch/model/prediction_task.py

@@ -46,9 +46,9 @@ def build(self, input_size) -> SequentialBlock:
 class BinaryClassificationTask(PredictionTask):
     DEFAULT_LOSS = torch.nn.BCELoss()
     DEFAULT_METRICS = (
-        tm.Precision(num_classes=2),
-        tm.Recall(num_classes=2),
-        tm.Accuracy(),
+        tm.Precision(num_classes=2, task="binary"),
+        tm.Recall(num_classes=2, task="binary"),
+        tm.Accuracy(task="binary"),
         # TODO: Fix this: tm.AUC()
     )