Open Source Text-To-Speech Projects: WhisperSpeech - In Depth Discussion

WhisperSpeech is a promising new open source TTS model that and be training on AUDIO ONLY data that already shows promising results after a few hundred GPU hours of training with a small 80M parameter model. In this video we talk about how it works and plans to scale it up to reach really high quality. Here is a voice sample of the 80M parameter model trained only a few days on 4 x 4090s: Here is the GitHub repo: If you wanna support this project join our LAION Discord Server check out our audio-generation channel: Check also out this realtime transcription project frm Collabora: ,

#