Update app.py

app.py

@@ -14,6 +14,10 @@ tts_tokenizer = AutoTokenizer.from_pretrained("Baghdad99/english_voice_tts")
 tts_model = AutoModelForTextToWaveform.from_pretrained("Baghdad99/english_voice_tts")
 
 def translate_speech(speech):
+    # Convert stereo to mono if necessary
+    if len(speech.shape) > 1:
+        speech = speech.mean(axis=0)
+
     # Transcribe the speech to text
     inputs = asr_processor(speech, return_tensors="pt", padding=True)
     logits = asr_model(inputs.input_values).logits