parambharat/kannada_asr_corpus
The corpus contains roughly 360 hours of audio and transcripts in Kannada language. The transcripts have beed de-duplicated using exact match deduplication.
029
fix dataloading script issue
fix broken mp3 files
add readme and loading script
add data files
initial commit