add comment

main.py

@@ -18,6 +18,7 @@ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 # comment in/out the models you want to use
 # RAM requirements: ~16GB x #models (+ ~4GB overhead)
 # VRAM requirements: ~16GB
+# if using int8: ~8GB VRAM x #models, low RAM requirements
 MODEL_CONFIGS = {
     "Llama-2-7b-chat-hf": {
         "identifier": "meta-llama/Llama-2-7b-chat-hf",