Hi everyone
A few weeks ago, we announced our plans to overhaul Common Voice localization. I’m pleased to share that the process is now complete, and you can find more details in our latest blog post here.
To summarize: The overhaul of Common Voice localization has significantly reduced the workload required for communities to start collecting data. Previously, localizing 824 strings across the entire user interface was required, but now only 300 core strings need to be localized. This change allows communities to begin data collection earlier and more efficiently, accommodating a range of data collection goals, from fine-tuning existing models to training new ones. The goal is to make Common Voice more accessible to languages with limited resources, particularly low-resourced languages.
Thanks
Gina