Fix typo in classification function selection logic to improve code c…

…onsistency (huggingface#32031)

Make problem_type condition consistent with num_labels condition

The latter condition generally overrides the former, so this is more of a code reading issue. I'm not sure the bug would ever actually get triggered under normal use.

src/transformers/pipelines/image_classification.py

@@ -171,9 +171,9 @@ def _forward(self, model_inputs):
 
     def postprocess(self, model_outputs, function_to_apply=None, top_k=5):
         if function_to_apply is None:
-            if self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels == 1:
+            if self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels == 1:
                 function_to_apply = ClassificationFunction.SIGMOID
-            elif self.model.config.problem_type == "single_label_classification" or self.model.config.num_labels > 1:
+            elif self.model.config.problem_type == "multi_label_classification" or self.model.config.num_labels > 1:
                 function_to_apply = ClassificationFunction.SOFTMAX
             elif hasattr(self.model.config, "function_to_apply") and function_to_apply is None:
                 function_to_apply = self.model.config.function_to_apply