fix: Make destructor to automatically call .close() method on Llama class.

Author: abetlen

llama_cpp/_internals.py

@@ -64,6 +64,9 @@ def free_model():
     def close(self):
         self._exit_stack.close()
 
+    def __del__(self):
+        self.close()
+
     def vocab_type(self) -> int:
         assert self.model is not None
         return llama_cpp.llama_vocab_type(self.model)
@@ -292,6 +295,9 @@ def free_ctx():
     def close(self):
         self._exit_stack.close()
 
+    def __del__(self):
+        self.close()
+
     def n_ctx(self) -> int:
         assert self.ctx is not None
         return llama_cpp.llama_n_ctx(self.ctx)
@@ -531,6 +537,9 @@ def free_batch():
     def close(self):
         self._exit_stack.close()
 
+    def __del__(self):
+        self.close()
+
     def n_tokens(self) -> int:
         assert self.batch is not None
         return self.batch.n_tokens

llama_cpp/llama.py

@@ -1968,6 +1968,9 @@ def close(self) -> None:
         """Explicitly free the model from memory."""
         self._stack.close()
 
+    def __del__(self) -> None:
+        self.close()
+
     @staticmethod
     def logits_to_logprobs(
         logits: Union[npt.NDArray[np.single], List], axis: int = -1