llm20

Some work to solve the Kaggle LLM 20 Questions

Submission

0. put model weights in submission/input folder

1. compress work

   tar -cf - -C submission . | pigz --fast | pv > submission.tar.gz

2. submit

   kaggle competitions submit -c llm-20-questions -f submission.tar.gz -m "<submission_message>"