The largest open instruction-tuning parallel corpus for Pashto ā English
š Why Pashto? Why Now?
Pashto (پŚŲŖŁ) is spoken by over 50 million people worldwide ā in Afghanistan, Pakistan, and the global diaspora. Yet it remains severely underrepresented in AI:
ā No Pashto-native LLM benchmarks
ā Limited high-quality parallel corpora
ā Most datasets are small, noisy, or not instruction-tuned