add code

app.py

@@ -39,7 +39,7 @@ def transcribe(audio: tuple[int, np.ndarray], conversation: list[dict], gradio_c
    
     output = pipe({"audio": audio_sr, "turns": conversation, "sampling_rate": target_sr},
                   max_new_tokens=512)
-    transcription = whisper({"array": audio_sr, "sampling_rate": target_sr})
+    transcription = whisper({"array": audio_sr.squeeze(), "sampling_rate": target_sr})
 
     conversation.append({"role": "user", "content": transcription})
     conversation.append({"role": "assistant", "content": output})