Skip to content

Speed up train completion usage serialization #1787

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

Merged
xeophon merged 1 commit into from
Jun 21, 2026
Merged

Speed up train completion usage serialization #1787

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
38 changes: 16 additions & 22 deletions verifiers/v1/clients/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,11 +13,6 @@
from typing import Any

from openai import AsyncOpenAI, OpenAIError
from openai.types import CompletionUsage
from openai.types.completion_usage import (
CompletionTokensDetails,
PromptTokensDetails,
)
from renderers import RenderedTokens
from renderers import OverlongPromptError as RendererOverlongPromptError
from renderers import RendererConfig
Expand Down Expand Up @@ -66,23 +61,22 @@ def serialize_completion(response: Response, model: str) -> dict:
}
for c in response.message.tool_calls
]
usage: CompletionUsage | None = None
usage: dict | None = None
if response.usage:
usage = CompletionUsage(
prompt_tokens=response.usage.input_tokens,
completion_tokens=response.usage.completion_tokens,
total_tokens=response.usage.total_tokens,
prompt_tokens_details=PromptTokensDetails(
cached_tokens=response.usage.cached_input_tokens
)
if response.usage.cached_input_tokens is not None
else None,
completion_tokens_details=CompletionTokensDetails(
reasoning_tokens=response.usage.reasoning_tokens
)
if response.usage.reasoning_tokens is not None
else None,
)
# Usage is validated earlier in the pipeline; building its wire dict directly saves time.
usage = {
"completion_tokens": response.usage.completion_tokens,
"prompt_tokens": response.usage.input_tokens,
"total_tokens": response.usage.total_tokens,
}
if response.usage.reasoning_tokens is not None:
usage["completion_tokens_details"] = {
"reasoning_tokens": response.usage.reasoning_tokens
}
if response.usage.cached_input_tokens is not None:
usage["prompt_tokens_details"] = {
"cached_tokens": response.usage.cached_input_tokens
}
return {
"id": response.id or "vf-intercept",
"object": "chat.completion",
Expand All @@ -95,7 +89,7 @@ def serialize_completion(response: Response, model: str) -> dict:
"finish_reason": response.finish_reason or "stop",
}
],
"usage": usage.model_dump(exclude_none=True) if usage is not None else None,
"usage": usage,
}


Expand Down
Loading