Adding OZ Eval task

[Stopwolf]

Add OZ Eval task for evaluating General Knowledge of LLMs in Serbian. More can be seen DjMel/oz-eval.

[clefourrier]

Hi!
Thanks for adding your evaluation.
A number of issues must be addressed - this PR opens up our code to a number of possible vulnerabilities, please change it.

[NathanHB]

Hi ! Thanks for your contribution. I tested the task locally and you need to change the null to None. Also you will need to format your code with make style for the tests to pass.

[NathanHB]

@clefourrier You need to accept the changes so that we can merge the PR :)

[clefourrier]

Tests are not passing though?

[Stopwolf]

Tests arent passing because of Invalid username or password in the following test:

tests/logging/test_evaluation_tracker.py:147: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
src/lighteval/logging/evaluation_tracker.py:185: in save
    self.push_to_hub(
src/lighteval/logging/evaluation_tracker.py:252: in push_to_hub
    self.api.create_repo(repo_id, private=not (self.public), repo_type="dataset", exist_ok=True)

Any idea how to fix this, or if it's even something on my side? Thanks @NathanHB 🤗

[clefourrier]

Hi! We normally added a fix for this (sorry for the late answer), re-running the tests again

[clefourrier]

Thanks for the eval!