Update app.py

app.py

@@ -13,7 +13,7 @@ tokenizer = AutoTokenizer.from_pretrained(model, token=token)
 llama_pipeline = pipeline(
     "text-generation",
     model=model,
-    torch_dtype = torch.float32,
+    torch_dtype = torch.bfloat16,
     device_map="auto"
 )