diff --git a/fastchat/model/model_adapter.py b/fastchat/model/model_adapter.py
index e7d14f918..975daa079 100644
--- a/fastchat/model/model_adapter.py
+++ b/fastchat/model/model_adapter.py
@@ -1681,6 +1681,31 @@ def get_default_conv_template(self, model_path: str) -> Conversation:
         return get_conv_template("llama2-chinese")
 
 
+class Lamma2ChineseAlpacaAdapter(BaseModelAdapter):
+    """The model adapter for ymcui/Chinese-LLaMA-Alpaca sft"""
+
+    def match(self, model_path: str):
+        return "chinese-alpaca" in model_path.lower()
+
+    def load_model(self, model_path: str, from_pretrained_kwargs: dict):
+        revision = from_pretrained_kwargs.get("revision", "main")
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_path,
+            trust_remote_code=True,
+            revision=revision,
+        )
+        model = AutoModelForCausalLM.from_pretrained(
+            model_path,
+            trust_remote_code=True,
+            low_cpu_mem_usage=True,
+            **from_pretrained_kwargs,
+        )
+        return model, tokenizer
+
+    def get_default_conv_template(self, model_path: str) -> Conversation:
+        return get_conv_template("llama2-chinese")
+
+
 class VigogneAdapter(BaseModelAdapter):
     """The model adapter for vigogne (e.g., bofenghuang/vigogne-2-7b-chat)"""
 
@@ -1883,6 +1908,7 @@ def get_default_conv_template(self, model_path: str) -> Conversation:
 register_model_adapter(BGEAdapter)
 register_model_adapter(E5Adapter)
 register_model_adapter(Lamma2ChineseAdapter)
+register_model_adapter(Lamma2ChineseAlpacaAdapter)
 register_model_adapter(VigogneAdapter)
 register_model_adapter(OpenLLaMaOpenInstructAdapter)
 register_model_adapter(ReaLMAdapter)
diff --git a/fastchat/model/model_registry.py b/fastchat/model/model_registry.py
index da08c2e26..840d91da4 100644
--- a/fastchat/model/model_registry.py
+++ b/fastchat/model/model_registry.py
@@ -330,6 +330,12 @@ def get_model_info(name: str) -> ModelInfo:
     "https://huggingface.co/FlagAlpha/Llama2-Chinese-13b-Chat",
     "Llama2-Chinese is a multi-language large-scale language model (LLM), developed by FlagAlpha.",
 )
+register_model_info(
+    ["Chinese-Alpaca-2-7B", "Chinese-Alpaca-2-13B"],
+    "Chinese-Alpaca",
+    "https://huggingface.co/hfl/chinese-alpaca-2-13b",
+    "New extended Chinese vocabulary beyond Llama-2, open-sourcing the Chinese LLaMA-2 and Alpaca-2 LLMs.",
+)
 register_model_info(
     ["Vigogne-2-7B-Instruct", "Vigogne-2-13B-Instruct"],
     "Vigogne-Instruct",