Update app.py

app.py

@@ -27,8 +27,10 @@ def transcribe(audio):
         resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
         audio_tensor = resampler(audio_tensor)
     
-    # Convert back to a NumPy array
+    # Convert back to a NumPy array and ensure it's in the correct shape
     audio_np = audio_tensor.cpu().numpy()
+    if audio_np.ndim == 2:
+        audio_np = audio_np
     
     pipe = pipeline(
         "automatic-speech-recognition",