Skip to content

fix(inspect): align GenerateConfig field names with current inspect-ai (fixes #1327) #1328

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

Open
sk8ordie84 wants to merge 2 commits into
base: main
Choose a base branch
from
+24 −5
Open

fix(inspect): align GenerateConfig field names with current inspect-ai (fixes #1327) #1328

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b12ac5b
fix(inspect): align GenerateConfig field names with current inspect-ai
sk8ordie84 Aug 12, 2026
b038d69
style: ruff blank-line fixes
sk8ordie84 Aug 12, 2026
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
29 changes: 24 additions & 5 deletions src/lighteval/main_inspect.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,25 @@
logger = logging.getLogger(__name__)


def _response_format_kwargs(response_format: str | None) -> dict:
"""inspect-ai removed GenerateConfig.response_format (>= 0.3.141).

Pass the flag through only when the installed inspect-ai accepts it, so a
fresh install does not crash at startup with a pydantic ValidationError.
"""
if response_format is None:
return {}
from inspect_ai.model import GenerateConfig

if "response_format" in GenerateConfig.model_fields:
return {"response_format": response_format}
logger.warning(
"--response-format is not supported by the installed inspect-ai (%s); ignoring it.",
getattr(__import__("inspect_ai"), "__version__", "unknown"),
)
return {}


@task
def get_inspect_ai_task(
lighteval_task_config: LightevalTaskConfig,
Expand Down Expand Up @@ -267,7 +286,7 @@ def eval( # noqa C901
rich_help_panel=HELP_PANEL_NAME_1,
),
] = None,
frequence_penalty: Annotated[
frequency_penalty: Annotated[
float | None,
Option(
help="Number between -2.0 and 2.0, Penalizes tokens that appear in the text too frequently, reducing repetition.",
Expand Down Expand Up @@ -309,7 +328,7 @@ def eval( # noqa C901
rich_help_panel=HELP_PANEL_NAME_1,
),
] = None,
log_probs: Annotated[
logprobs: Annotated[
bool | None,
Option(
help="Returns log probabilities for each token in the generated text", rich_help_panel=HELP_PANEL_NAME_1
Expand Down Expand Up @@ -474,19 +493,19 @@ def eval( # noqa C901
temperature=temperature,
top_p=top_p,
top_k=top_k,
frequence_penalty=frequence_penalty,
frequency_penalty=frequency_penalty,
presence_penalty=presence_penalty,
seed=seed,
stop_seqs=stop_seqs,
num_choices=num_choices,
best_of=best_of,
log_probs=log_probs,
logprobs=logprobs,
top_logprobs=top_logprobs,
cache_prompt=cache_prompt,
reasoning_effort=reasoning_effort,
reasoning_tokens=reasoning_tokens,
reasoning_history=reasoning_history,
response_format=response_format,
**_response_format_kwargs(response_format),
parallel_tool_calls=parallel_tool_calls,
max_tool_output=max_tool_output,
internal_tools=internal_tools,
Expand Down