Open Internet by MindsNet

Implementing Effective Text-to-Speech Synthesis with Limited Data

The author is struggling to implement the Pocket TTS model, facing issues with generating meaningful text and voice cloning. Despite trying various techniques, such as scheduled sampling and adding noise, the model fails to produce satisfactory results. The author seeks suggestions on how to improve the implementation.

Computing & Technology, Computer Science, Artificial Intelligence