Update app.py

app.py

@@ -77,10 +77,12 @@ def transcribe_and_translate(audio_input):
         return_dict=True,
         return_tensors='pt'
     ).to('cuda')
+
+    streamer = TextStreamer(processor, skip_prompt = True)
     generation_kwargs = dict(
         **inputs,
         temperature = 1.0, top_p = 0.95, top_k = 64,
-        streamer = TextStreamer(processor, skip_prompt = True),
+        streamer = streamer,
         max_new_tokens=1024,
     )
     thread = Thread(target=model.generate, kwargs=generation_kwargs)