Pashto Common Voice: Building the First Open Speech Corpus for a 60-Million-Speaker Low-Resource Language
The Pashto Common Voice corpus is a large-scale, open speech resource built by a community between 2022 and 2025 to address Pashto's lack of open speech tech. It grew from minimal starting points to over 147 hours and 60,337 validated clips. This resource allows for the fine-tuning of ASR systems, proving that full fine-tuning is possible even for low-resource languages.