diff --git a/onnxruntime/python/tools/transformers/convert_generation.py b/onnxruntime/python/tools/transformers/convert_generation.py
index 596c556cfc..8c3ecceeff 100644
--- a/onnxruntime/python/tools/transformers/convert_generation.py
+++ b/onnxruntime/python/tools/transformers/convert_generation.py
@@ -363,6 +363,22 @@ def parse_arguments(argv: Optional[List[str]] = None) -> argparse.Namespace:
         help="Vocab_size of the underlying model used to decide the shape of vocab mask",
     )
 
+    beam_parameters_group.add_argument(
+        "--eos_token_id",
+        type=int,
+        required=False,
+        default=-1,
+        help="custom eos_token_id for generating model with existing onnx encoder/decoder",
+    )
+
+    beam_parameters_group.add_argument(
+        "--pad_token_id",
+        type=int,
+        required=False,
+        default=-1,
+        help="custom pad_token_id for generating model with existing onnx encoder/decoder",
+    )
+
     test_group = parser.add_argument_group("Other options for testing parity and performance")
 
     test_group.add_argument(
@@ -1374,6 +1390,11 @@ def convert_generation_model(args: argparse.Namespace, generation_type: Generati
     if args.vocab_size != -1:
         vocab_size = args.vocab_size
 
+    if args.eos_token_id != -1:
+        eos_token_id = args.eos_token_id
+    if args.pad_token_id != -1:
+        pad_token_id = args.pad_token_id
+
     decoder_model = onnx.load_model(args.decoder_onnx, load_external_data=True)
     decoder_model.graph.name = f"{args.model_type} decoder"