Skip to content

RiTA-nlp/ita-eval

 
 

Repository files navigation

ItaEval

Webpage Leaderboard Paper

This repository contains the configuration and code utilities to run the ItaEval evaluation suite.

Exploring the Suite

All the configuration file are under lm_eval/tasks/ita_eval. We also have included it as a "benchmark" under lm_eval/tasks/benchmarks.

Getting Started

We release several runner bash scripts to run base and chat models against the suite. Head to bash/ to find them.

Note that the recipes listed in the folder are tailored to our hardware and you will very likely need to adapt them to yours.

Run your own model

In a scenario where all of the dependencies are installed correctly, you should be able to run your model on ItaEval with

MODEL="your-model-id-on-the-huggingface-hub"
lm_eval --model hf \
    --model_args pretrained=${MODEL},dtype=bfloat16 \
    --tasks ita_eval \
    --batch_size 1 \
    --log_samples \
    --output_path "."

Add a model to the Leaderboard

Follow these steps:

  1. Run the evaluation with the code above. You will end up with a folder containing a file starting with results_
  2. Copy and push that folder into this directory: https://huggingface.co/datasets/RiTA-nlp/ita-eval-results/
  3. Edit the model_info.yaml file to add the information about the new model(s)
  4. Run this script from the main directory of the ita-eval-results repository.
  5. Push the changes.

Note, points 2 through 5 require having access to the results repository.

Acknowledgments

ItaEval and TweetyIta are the results of the joint effort of members of the Risorse per la Lingua Italiana community. We thank every member that dedicated their personal time to the sprints. We thank CINECA for providing the computational resources (ISCRA grant: HP10C3RW9F).

Cite

@inproceedings{attanasio2024itaeval,
  title={ItaEval and TweetyIta: A New Extensive Benchmark and Efficiency-First Language Model for Italian},
  author={Attanasio, Giuseppe and Delobelle, Pieter and La Quatra, Moreno and Santilli, Andrea and Savoldi, Beatrice},
  booktitle={CLiC-it 2024: Tenth Italian Conference on Computational Linguistics, Date: 2024/12/04-2024/12/06, Location: Pisa, Italy},
  year={2024}
}