Update app.py

app.py

@@ -46,7 +46,7 @@ def transcribe(audio: tuple[int, np.ndarray], transformers_convo: list[dict], gr
     name = str(uuid.uuid4()) + ".mp3"
     segment.export(name, format="mp3")
     transformers_convo.append({"role": "user", "content": [{"type": "audio", "audio_url": name}]})
-    gradio_convo.append({"role": "assistant", "content": gr.Audio(value=name)})
+    gradio_convo.append({"role": "user", "content": gr.Audio(value=name)})
     text = processor.apply_chat_template(transformers_convo, add_generation_prompt=True, tokenize=False)
     audios = []
     for message in transformers_convo: