diff --git a/quantize_modelopt.py b/quantize_modelopt.py
index d3c5de1..ad9f801 100644
--- a/quantize_modelopt.py
+++ b/quantize_modelopt.py
@@ -100,6 +100,10 @@ def main():
     }
     if args.use_seq_device_map:
         model_kwargs["device_map"] = "auto"
+        model_kwargs["offload_folder"] = "offload"
+        model_kwargs["offload_state_dict"] = True
+        model_kwargs["max_memory"] = {i: "160GiB" for i in range(8)}
+        model_kwargs["max_memory"]["cpu"] = "2500GiB"
 
     model = AutoModelForCausalLM.from_pretrained(args.model, **model_kwargs)