fb-mms-tts-pan-rajinder_singh_corrected-epochs-15-finetuned

A fine-tuned Punjabi text-to-speech model derived from Meta's MMS TTS Punjabi Eastern checkpoint, trained for 15 epochs on corrected data to synthesize spoken Punjabi audio in a specific learned voice. It uses the VITS-style architecture from Meta's Massively Multilingual Speech project. Source: supabase.