one more losses fix

spacy/pipeline/tagger.pyx

@@ -227,10 +227,13 @@ class Tagger(TrainablePipe):
 
         DOCS: https://nightly.spacy.io/api/tagger#rehearse
         """
+        if losses is None:
+            losses = {}
+        losses.setdefault(self.name, 0.0)
         validate_examples(examples, "Tagger.rehearse")
         docs = [eg.predicted for eg in examples]
         if self._rehearsal_model is None:
-            return
+            return losses
         if not any(len(doc) for doc in docs):
             # Handle cases where there are no tokens in any docs.
             return losses
@@ -240,8 +243,6 @@ class Tagger(TrainablePipe):
         gradient = guesses - target
         backprop(gradient)
         self.finish_update(sgd)
-        if losses is not None:
-            losses.setdefault(self.name, 0.0)
         losses[self.name] += (gradient**2).sum()
         return losses