Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix BaseOpenAIChatCompletionClient token usage #4770

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
+9 −11
Open

Fix BaseOpenAIChatCompletionClient token usage #4770

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
f9cf23a
Fix openai client token usage (#4769)
gziz Dec 20, 2024
0136055
Include actual_usage in add_usage function
gziz Dec 26, 2024
813029c
Merge branch 'main' into fix_client_token_track
gziz Dec 26, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 9 additions & 11 deletions python/packages/autogen-ext/src/autogen_ext/models/openai/_openai_client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -293,13 +293,6 @@ def calculate_vision_tokens(image: Image, detail: str = "auto") -> int:
return total_tokens


def _add_usage(usage1: RequestUsage, usage2: RequestUsage) -> RequestUsage:
return RequestUsage(
prompt_tokens=usage1.prompt_tokens + usage2.prompt_tokens,
completion_tokens=usage1.completion_tokens + usage2.completion_tokens,
)


def convert_tools(
tools: Sequence[Tool | ToolSchema],
) -> List[ChatCompletionToolParam]:
Expand Down Expand Up @@ -572,8 +565,7 @@ async def create(
logprobs=logprobs,
)

_add_usage(self._actual_usage, usage)
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm wondering what's the difference between these two...

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Traditionally actual vs total is a matter of if the tokens were cached or not. But since we don't have caching atm they should be identical

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @jackgerrits! I have updated the PR to use both actual_usage and total_usage.

_add_usage(self._total_usage, usage)
self.add_usage(usage)

# TODO - why is this cast needed?
return response
Expand Down Expand Up @@ -789,8 +781,7 @@ async def create_stream(
logprobs=logprobs,
)

_add_usage(self._actual_usage, usage)
_add_usage(self._total_usage, usage)
self.add_usage(usage)

yield result

Expand Down Expand Up @@ -897,6 +888,13 @@ def remaining_tokens(self, messages: Sequence[LLMMessage], tools: Sequence[Tool
def capabilities(self) -> ModelCapabilities:
return self._model_capabilities

def add_usage(self, usage: RequestUsage) -> None:
self._total_usage.prompt_tokens += usage.prompt_tokens
self._total_usage.completion_tokens += usage.completion_tokens

self._actual_usage.prompt_tokens += usage.prompt_tokens
self._actual_usage.completion_tokens += usage.completion_tokens


class OpenAIChatCompletionClient(BaseOpenAIChatCompletionClient, Component[OpenAIClientConfigurationConfigModel]):
"""Chat completion client for OpenAI hosted models.
Expand Down
Loading