Update app.py

app.py

@@ -34,6 +34,7 @@ def parse(wav_file):
     input_values = read_file_and_process(wav_file)
     with torch.no_grad():
         logits = model(**input_values).logits
+    return parse_transcription(logits)