diff --git a/src/llmtuner/tuner/core/loader.py b/src/llmtuner/tuner/core/loader.py
index 921dbc11..22b657e3 100644
--- a/src/llmtuner/tuner/core/loader.py
+++ b/src/llmtuner/tuner/core/loader.py
@@ -18,10 +18,11 @@ from trl import AutoModelForCausalLMWithValueHead
 from llmtuner.extras.logging import get_logger
 from llmtuner.extras.misc import count_parameters, prepare_model_for_training
 from llmtuner.extras.save_and_load import load_valuehead_params
+from llmtuner.hparams import FinetuningArguments
 from llmtuner.tuner.core.adapter import init_adapter
 
 if TYPE_CHECKING:
-    from llmtuner.hparams import ModelArguments, FinetuningArguments
+    from llmtuner.hparams import ModelArguments
 
 
 logger = get_logger(__name__)
diff --git a/src/llmtuner/tuner/core/trainer.py b/src/llmtuner/tuner/core/trainer.py
index 928b0d9b..805a3553 100644
--- a/src/llmtuner/tuner/core/trainer.py
+++ b/src/llmtuner/tuner/core/trainer.py
@@ -68,7 +68,7 @@ class PeftTrainer(Seq2SeqTrainer):
         else:
             torch.save(state_dict, os.path.join(output_dir, WEIGHTS_NAME))
 
-        if self.tokenizer is not None:
+        if self.finetuning_args.finetuning_type == "full" and self.tokenizer is not None:
             self.tokenizer.save_pretrained(output_dir)
 
         with open(os.path.join(output_dir, TRAINING_ARGS_NAME), "w", encoding="utf-8") as f: