8000 [BUG] evaluation on minervamath · Issue #628 · huggingface/lighteval · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
[BUG] evaluation on minervamath #628
Closed
@wanghaoyu0408

Description

@wanghaoyu0408

I try to evaluate my model on math-ai/minervamath. I used LatexExtractionConfig() in the following way:

latex_gold_metric = multilingual_extractive_match_metric( language=Language.ENGLISH, fallback_mode="first_match", precision=5, gold_extraction_target=(LatexExtractionConfig(),), # Match boxed first before trying other regexes pred_extraction_target=(ExprExtractionConfig(), LatexExtractionConfig(boxed_match_priority=0)), aggregation_function=max, )

def minervamath_prompt_fn(line, task_name: str = None): return Doc( task_name=task_name, query=MATH_QUERY_TEMPLATE.format(Question=line["question"]), choices=[line["answer"]], gold_index=0, )

minervamath = LightevalTaskConfig( name="minervamath", suite=["custom"], prompt_function=minervamath_prompt_fn, hf_repo="math-ai/minervamath", hf_subset="default", hf_avail_splits=["test"], evaluation_splits=["test"], few_shots_split=None, few_shots_select=None, generation_size=32768, metric=[latex_gold_metric], version=1, )

But I have following warning:

Image

How to fix it?

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions

      0