Update processing_aero.py

processing_aero.py

@@ -37,7 +37,6 @@ class AeroProcessorKwargs(ProcessingKwargs, total=False):
             "padding": False,
         },
         "audio_kwargs": {
-            "return_tensors": "pt"
         },
     }
 
@@ -116,7 +115,9 @@ class AeroProcessor(ProcessorMixin):
             tokenizer_init_kwargs=self.tokenizer.init_kwargs,
             **kwargs,
         )
-        output_kwargs["audio_kwargs"].pop("padding")
+        # Make sure no duplicate padding
+        if "padding" in output_kwargs["audio_kwargs"]:
+            output_kwargs["audio_kwargs"].pop("padding")
 
         if isinstance(text, str):
             text = [text]