Update app.py

app.py

@@ -13,7 +13,7 @@ print("CPU threads:", torch.get_num_threads())
 
 
 if torch_device == "cuda":
-    model = AutoModelForCausalLM.from_pretrained(model_id, load_in_8bit=True, device_map="auto")
+    model = AutoModelForCausalLM.from_pretrained(model_id, load_in_4bit=True, device_map="auto")
 else:
     model = AutoModelForCausalLM.from_pretrained(model_id)
 tokenizer = AutoTokenizer.from_pretrained(model_id)