use tinyroberta for qa

app.py

@@ -24,8 +24,8 @@ model = whisper.load_model("base", device = device)
 model_med = whisper.load_model("medium", device = device)
 
 #Roberta Q&A
-#model_name = "deepset/tinyroberta-squad2"
-model_name = "deepset/roberta-base-squad2"
+model_name = "deepset/tinyroberta-squad2"
+#model_name = "deepset/roberta-base-squad2"
 nlp = pipeline('question-answering', model=model_name, tokenizer=model_name, device = 0 if device == "cuda" else -1)
 
 #TTS