From 7e4e6782d1651dd77d278181073655f973874153 Mon Sep 17 00:00:00 2001 From: Paul Gauthier Date: Fri, 27 Sep 2024 16:46:03 -0700 Subject: [PATCH] fix: set max_tokens to 1 for cache warming --- aider/coders/base_coder.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/aider/coders/base_coder.py b/aider/coders/base_coder.py index 5537bb8ed..55bf2b32f 100755 --- a/aider/coders/base_coder.py +++ b/aider/coders/base_coder.py @@ -1071,14 +1071,15 @@ class Coder: self.warming_pings_left -= 1 self.next_cache_warm = time.time() + delay + kwargs = self.main_model.extra_params or dict() + kwargs["max_tokens"] = 1 + try: completion = litellm.completion( model=self.main_model.name, messages=self.cache_warming_chunks.cacheable_messages(), stream=False, - max_tokens=1, - extra_headers=self.main_model.extra_headers, - **self.main_model.extra_params, # Use **kwargs here + **kwargs, ) except Exception as err: self.io.tool_warning(f"Cache warming error: {str(err)}")