Update model.py

model.py

@@ -11,10 +11,10 @@ def load_model(
 
     model = LlamaForCausalLM.from_pretrained(
         base,
-        # load_in_8bit=True,
+        load_in_8bit=True,
         device_map="auto",
     )
     
-    # model = PeftModel.from_pretrained(model, finetuned, device_map={'': 0})
+    model = PeftModel.from_pretrained(model, finetuned, device_map={'': 0})
     return model, tokenizer