This repository contains the necessary files and instructions for SemEval2025-Task11 participants to check the formatting of their submissions and evaluate them.
Install the required dependencies by running:
pip install -r requirements.txt
Ensure that your submission file is properly structured and contains the required predictions using the language code as per the task. The languages are:
afr, arq, amh, arb, ary, eng, esp, deu, hau, hin, ibo, ind, jav, kin, mar, orm, ptbr, ptm, pcm, ron, rus, som, sun, swa, swe, tat, tir, tsn, tso, ukr, xho, yor, zho, zul
To ensure your submission is valid and meets the requirements, use the check_submission.py
script. This will ensure that your file follows the task’s requirements. Check this guide for more information about participation and submission.
python check_submission.py -p dev -s <path_to_your_submission_file>
This script checks whether:
- the structure and the formatting are correct,
- the necessary predictions (per language) are included,
- the files comply with the shared task’s requirements in general.
Make sure that this script runs without errors before proceeding with the evaluation.
- Examples of properly formatted files: Track A (English) and Track B (German))
Once the submission passes the format validation, you can evaluate its performance by running the evaluation script provided in the Jupyter notebook check_submission.ipynb
.
- Open the
check_submission.ipynb
file in a Jupyter environment. - Make sure to set the correct file paths for your submission and test data.
- Run the cells to compute the evaluation metrics (e.g., accuracy, precision, recall, F1 scores).
Alternatively, if you're running it as a script, make sure thta you have the necessary parameters (e.g., the right paths to your files).
After the evaluation, the script will output various metrics (see below, example for German), such as accuracy, precision, recall, and F1 scores for each language.
Note This is an example for Track A.
- If you want to automatically zip the file after validation and evaluation, add this flag (see below):
--zip_for_submission \
- The script zips the submission file automatically and your file will be ready to upload! The following message will confirm that your file is zipped:
Zipped file: pred_<language_code>_<task>.zip is ready for upload in the Codalab submission page.
- Upload the zipped file to the Codabench as your final submission.