diff --git a/llama_cpp/server/errors.py b/llama_cpp/server/errors.py index febe3e3..9d3d355 100644 --- a/llama_cpp/server/errors.py +++ b/llama_cpp/server/errors.py @@ -72,7 +72,7 @@ class ErrorResponseFormatters: return 400, ErrorResponse( message=message.format( context_window, - completion_tokens + prompt_tokens, + (completion_tokens or 0) + prompt_tokens, prompt_tokens, completion_tokens, ), # type: ignore