Update app.py

app.py

@@ -20,7 +20,7 @@ whisper_model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-
 # Initialize the summarization model and tokenizer
 # Use a smaller version of the Llama model and load in FP16
 summarization_model = AutoModelForCausalLM.from_pretrained(
-    "meta-llama/Llama-2-3b-hf",
+    "meta-llama/LlamaGuard-7b",
     torch_dtype=torch.float16,
     low_cpu_mem_usage=True
 )