diff --git a/src/llmtuner/data/template.py b/src/llmtuner/data/template.py
index d4ec22d5..ee7c0da1 100644
--- a/src/llmtuner/data/template.py
+++ b/src/llmtuner/data/template.py
@@ -758,6 +758,26 @@ register_template(
 )
 
 
+register_template(
+    name="yuan",
+    prefix=[
+        "{{system}}"
+    ],
+    prompt=[
+        "{{query}}",
+        {"token": "<sep>"}
+    ],
+    system="",
+    sep=[
+        "\n"
+    ],
+    stop_words=[
+        "<eod>"
+    ],
+    replace_eos=True
+)
+
+
 register_template(
     name="zephyr",
     prefix=[
diff --git a/src/llmtuner/hparams/model_args.py b/src/llmtuner/hparams/model_args.py
index f4e55d43..3f415bee 100644
--- a/src/llmtuner/hparams/model_args.py
+++ b/src/llmtuner/hparams/model_args.py
@@ -102,6 +102,10 @@ class ModelArguments:
         default=1024,
         metadata={"help": "The maximum length of the model inputs used for quantization."}
     )
+    export_lecacy_format: Optional[bool] = field(
+        default=False,
+        metadata={"help": "Whether or not to save the `.bin` files instead of `.safetensors`."}
+    )
 
     def __post_init__(self):
         self.compute_dtype = None
diff --git a/src/llmtuner/train/tuner.py b/src/llmtuner/train/tuner.py
index 0af124a3..033c20f5 100644
--- a/src/llmtuner/train/tuner.py
+++ b/src/llmtuner/train/tuner.py
@@ -47,7 +47,11 @@ def export_model(args: Optional[Dict[str, Any]] = None):
 
     model.config.use_cache = True
     model = model.to("cpu")
-    model.save_pretrained(model_args.export_dir, max_shard_size="{}GB".format(model_args.export_size))
+    model.save_pretrained(
+        save_directory=model_args.export_dir,
+        max_shard_size="{}GB".format(model_args.export_size),
+        safe_serialization=(not model_args.export_lecacy_format)
+    )
 
     try:
         tokenizer.padding_side = "left" # restore padding side