huggingface / lighteval

Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends
MIT License
845 stars 100 forks source link

Selecting tasks using their superset #308

Closed hynky1999 closed 1 month ago

hynky1999 commented 2 months ago

What does this implement/fix? Explain your changes.

clefourrier commented 2 months ago

You'll need to fix the tests first :)

clefourrier commented 1 month ago

@hynky1999 where are we wrt this PR? is it still active?

hynky1999 commented 1 month ago

Yes, just have other things to focus on rn. Will try to update it by the of this week

clefourrier commented 1 month ago

No urgency, just going through to see what to close

clefourrier commented 1 month ago

I skimmed it, that looks good, merging. Thanks for the update!