Free model when llama is unloaded. Closes #434

Author: abetlen

llama_cpp/llama.py

@@ -1437,6 +1437,9 @@ def create_chat_completion(
             return self._convert_text_completion_to_chat(completion)
 
     def __del__(self):
+        if self.model is not None:
+            llama_cpp.llama_free_model(self.model)
+            self.model = None
         if self.ctx is not None:
             llama_cpp.llama_free(self.ctx)
             self.ctx = None