scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-10 22:41:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 83

fix(bedrock/converse): capture raw input_tokens as text_tokens before cache inflation in PromptTokensDetailsWrapper

Browse source
This commit is contained in:
Ishaan Jaffer 2026-04-14 18:20:34 -07:00
parent 5c056cae9f
commit c84597ecd0
No known key found for this signature in database
1 changed files with 4 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
litellm/llms/bedrock/chat/converse_transformation.py
View file

@ -1651,6 +1651,7 @@ class AmazonConverseConfig(BaseConfig):
cache_creation_input_tokens: int = 0
cache_read_input_tokens: int = 0
raw_input_tokens = input_tokens # capture before inflation
if "cacheReadInputTokens" in usage:
cache_read_input_tokens = usage["cacheReadInputTokens"]
input_tokens += cache_read_input_tokens
@ -1659,7 +1660,9 @@ class AmazonConverseConfig(BaseConfig):
input_tokens += cache_creation_input_tokens
prompt_tokens_details = PromptTokensDetailsWrapper(
cached_tokens=cache_read_input_tokens
cached_tokens=cache_read_input_tokens,
cache_creation_tokens=cache_creation_input_tokens,
text_tokens=raw_input_tokens,
)
reasoning_tokens = (
token_counter(text=reasoning_content, count_response_tokens=True)