AI Resonance · 阅读最新日报 · AI 入门推荐

2026-11-18 · America/Los_Angeles · 行业新闻 · #6

A 4.5M-parameter TTS model runs 70x real-time on CPU and cost just $10 to train

tiny-voca is a small English text-to-speech model with two voices, F2 (female) and M2 (male). The complete model, including the waveform generator, has 4.5M parameters and was trained from scratch on about 18 hours of synthetic speech. It runs about 70x faster than real time on four CPU threads. Text of any length works. The command reads it one sentence at a time and joins the audio. A text that ends without a mark gets a full stop. A sentence with nothing to read, such as one in another script, is skipped. The saved audio never peaks above 0.99. The graph takes phoneme ids and stress values

A 4.5M-parameter TTS model runs 70x real-time on CPU and cost just $10 to train

热度 39.5 / 100;排名与评分保留该期记录。