Fish Audio S2 Pro

### TL;DR

Fish Audio S2 Pro is an advanced text-to-speech (TTS) model developed by Fish Audio, offering fine-grained inline control over prosody and emotion. Trained on over 10 million hours of audio data across more than 80 languages, it combines reinforcement learning alignment with a dual-autoregressive architecture to deliver high-quality, expressive speech synthesis.

Key Insights & Metrics

Key Features

• Fine-grained inline control over prosody and emotion
• Supports over 80 languages
• Dual-autoregressive architecture for efficient inference
• Real-time streaming performance on NVIDIA H200 GPU
• Open-source model weights and fine-tuning code

→ Related Releases

Discussion