Selective Classifier-free Guidance for Zero-shot Text-to-speech

1. 1Separate conditions for classifier-free guidance to enable trade-offs between speaker fidelity and text content adherence
2. 2Evaluate the effectiveness of selective classifier-free guidance in zero-shot text-to-speech
3. 3Apply the approach to speech synthesis to improve the balance between desired characteristics

ML researchers and engineers working on text-to-speech systems can benefit from this approach to improve the quality of their models, and software engineers can apply these findings to develop more efficient speech synthesis algorithms

💡 Selective classifier-free guidance can improve the balance between speaker fidelity and text content adherence in zero-shot text-to-speech

💡 Improving zero-shot text-to-speech with selective classifier-free guidance