VMLU submission

This repository was forked from https://github.com/ZaloAI-Jaist/VMLU. Please follow the below instructions to generate the submission file:

$ pip install -r requirements.txt
$ python test_prompt.py \
    --llm vtrungnhan9/vmlu-llm \
    --device cuda

After running these commands, the submission file: vtrungnhan9-vmlu-llm.csv is generated in ./logs folder. Please note that our model was publicly pushed to HuggingFace and it is automatically downloaded when you run test_prompt.py script.