johneortega/IWSLT2026_Quechua_data

Main repository for the sharing of Quechua-Spanish Speech Translation data as part of the low-resource shared task at IWSLT 2026 # IWSLT2026 - Low-resource Speech Translation Track: Quechua-Spanish Parallel Corpus Main repository for the sharing of Quechua-Spanish Speech Translation data as part of the low-resource shared task at IWSLT 2026. ## Data for the `unconstrained` task In addition to the 1 hour, 40 minutes of Quechua audio data aligned with Spanish translations, we also provided participants with a corpus of 48 hours of fully transcribed Quechua audio `without` translations for the `unconstrained` task. The audio data and corresponding transcriptions are a bigger extract from the Siminchik data set. The hope is that this data can be directly used for assistance in the development of speech recognition components for the `unconstrained` task. The data can be easily downloaded directly fron here: Unconstrained QUE-SPA Additional Audio 1. **Please Note**: Participants are not required to use this data but are free to use with the license below. ## Evaluation You can evaluate your model using two widely adopted translation metrics: BLEU and ChrF++. Follow this link link for detailed instructions on how to use the evaluation script. ## Citation ``` @article{cardenas2018siminchik, title={Siminchik: A speech corpus for preservation of southern quechua}, author={Cardenas, Ronald and Zevallos, Rodolfo and Baquerizo, Reynaldo and Camacho, Luis}, journal={ISI-NLP 2}, pages={21}, year={2018} } ``` ## Additional data `with translations` for the `unconstrained` task We are also sharing a secondary dataset named `que_spa_synthetic_translation`, which contains a set of synthetic Quechua-Spanish translations. This dataset is intended as additional training data for the unconstrained task. While these translations are machine-generated (translated by Google) and post-edited by a Quechua speaker, they can be valuable for experimenting with training strategies in the unconstrained setup. The dataset can be found in the `que_spa_synthetic_translation` folder, and participants are encou …