phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-24 18:24:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

(fix) trim_messages util docstring

Browse source
This commit is contained in:
ishaan-jaff 2023-10-27 08:35:48 -07:00
parent 895cb5d0f9
commit 1aa53f6322
1 changed files with 0 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

1
litellm/utils.py
View file

@ -4203,7 +4203,6 @@ def trim_messages(
Args:
messages: Input messages to be trimmed. Each message is a dictionary with 'role' and 'content'.
model: The LiteLLM model being used (determines the token limit).
system_message: Optional system message to preserve at the start of the conversation.
trim_ratio: Target ratio of tokens to use after trimming. Default is 0.75, meaning it will trim messages so they use about 75% of the model's token limit.
return_response_tokens: If True, also return the number of tokens left available for the response after trimming.
max_tokens: Instead of specifying a model or trim_ratio, you can specify this directly.