diff --git a/llama/generation.py b/llama/generation.py
index 9045c2f..5f8faf9 100755
--- a/llama/generation.py
+++ b/llama/generation.py
@@ -56,6 +56,7 @@ class Llama:
         max_seq_len: int,
         max_batch_size: int,
         model_parallel_size: Optional[int] = None,
+        seed: int = 1,
     ) -> "Llama":
         """
         Build a Llama instance by initializing and loading a pre-trained model.
@@ -91,7 +92,7 @@ class Llama:
         torch.cuda.set_device(local_rank)
 
         # seed must be the same in all processes
-        torch.manual_seed(1)
+        torch.manual_seed(seed)
 
         if local_rank > 0:
             sys.stdout = open(os.devnull, "w")