Update Submodule vendor/llama.cpp 1191758..3fc4e10

- Append `n_layer_nextn` call method

Signed-off-by: JamePeng <jame_peng@sina.com>

Author: JamePeng

llama_cpp/_internals.py

@@ -123,6 +123,9 @@ def n_embd_out(self) -> int:
     def n_layer(self) -> int:
         return llama_cpp.llama_model_n_layer(self.model)
 
+    def n_layer_nextn(self) -> int:
+        return llama_cpp.llama_model_n_layer_nextn(self.model)
+
     def n_head(self) -> int:
         return llama_cpp.llama_model_n_head(self.model)
 

llama_cpp/llama.py

@@ -3470,6 +3470,10 @@ def n_layer(self) -> int:
         """Return the n_layer value."""
         return self._model.n_layer()
 
+    def n_layer_nextn(self) -> int:
+        """Return the n_layer_nextn value."""
+        return self._model.n_layer_nextn()
+
     def n_head(self) -> int:
         """Return the head size."""
         return self._model.n_head()

llama_cpp/llama_cpp.py

@@ -1746,6 +1746,11 @@ def llama_model_n_layer(model: llama_model_p, /) -> int:
     ...
 
 
+# LLAMA_API int32_t llama_model_n_layer_nextn(const struct llama_model * model);
+@ctypes_function("llama_model_n_layer_nextn", [llama_model_p_ctypes], ctypes.c_int32)
+def llama_model_n_layer_nextn(model: llama_model_p, /) -> int:
+    ...
+
 # LLAMA_API int32_t llama_model_n_head     (const struct llama_model * model);
 @ctypes_function("llama_model_n_head", [llama_model_p_ctypes], ctypes.c_int32)
 def llama_model_n_head(model: llama_model_p, /) -> int: