Update handler.py

handler.py

@@ -32,7 +32,7 @@ class EndpointHandler:
         if audio is not None:
             audio_list = json.loads(audio)
             audio_array = np.array(audio_list)
-            parameters["audio"] = audio_array
+            audio = audio_array
 
         if duration is not None:
             # Calculate max new tokens based on duration, this is a placeholder, replace with actual logic
@@ -44,11 +44,14 @@ class EndpointHandler:
         inputs = self.processor(
             text=[inputs],
             padding=True,
-            return_tensors="pt",).to(self.device)
+            return_tensors="pt",
+            audio=audio).to(self.device)
 
         # If 'duration' is inside 'parameters', remove it
         if parameters is not None and 'duration' in parameters:
             parameters.pop('duration')
+        if parameters is not None and 'audio' in parameters:
+            parameters.pop('audio')
 
         # pass inputs with all kwargs in data
         if parameters is not None: