diff --git a/src/llamafactory/chat/hf_engine.py b/src/llamafactory/chat/hf_engine.py
index 5a5c00c8..61f7f9a6 100644
--- a/src/llamafactory/chat/hf_engine.py
+++ b/src/llamafactory/chat/hf_engine.py
@@ -171,7 +171,7 @@ class HuggingfaceEngine(BaseEngine):
             elif not isinstance(value, torch.Tensor):
                 value = torch.tensor(value)
 
-            gen_kwargs[key] = value.to(model.device)
+            gen_kwargs[key] = value.to(dtype=model.dtype, device=model.device)
 
         return gen_kwargs, prompt_length
 
diff --git a/src/llamafactory/webui/locales.py b/src/llamafactory/webui/locales.py
index 8b5baade..64cdf1f5 100644
--- a/src/llamafactory/webui/locales.py
+++ b/src/llamafactory/webui/locales.py
@@ -1119,7 +1119,7 @@ LOCALES = {
             "info": "Нормализация оценок в тренировке PPO.",
         },
         "zh": {
-            "label": "奖励模型",
+            "label": "归一化分数",
             "info": "PPO 训练中归一化奖励分数。",
         },
         "ko": {