On this auspicious day of Hayagreeva Jayanti, I am open-sourcing the data and models for Samskrutam-Chant Speech-to-Text (ASR), named सुश्रोता. ~ 18 h of utterance-level annotated Shastric text,…

Meet Su-Śrota - an open-source ASR built specifically for Sanskrit chanting. Fine-tuned on hundreds of hours of curated chant recordings, it listens to recitation, not just speech. Powering…