English

Accelerating Flow-Matching-Based Text-to-Speech via Empirically Pruned Step Sampling

Audio and Speech Processing 2025-06-05 v2 Sound

Abstract

Flow-matching-based text-to-speech (TTS) models, such as Voicebox, E2 TTS, and F5-TTS, have attracted significant attention in recent years. These models require multiple sampling steps to reconstruct speech from noise, making inference speed a key challenge. Reducing the number of sampling steps can greatly improve inference efficiency. To this end, we introduce Fast F5-TTS, a training-free approach to accelerate the inference of flow-matching-based TTS models. By inspecting the sampling trajectory of F5-TTS, we identify redundant steps and propose Empirically Pruned Step Sampling (EPSS), a non-uniform time-step sampling strategy that effectively reduces the number of sampling steps. Our approach achieves a 7-step generation with an inference RTF of 0.030 on an NVIDIA RTX 3090 GPU, making it 4 times faster than the original F5-TTS while maintaining comparable performance. Furthermore, EPSS performs well on E2 TTS models, demonstrating its strong generalization ability.

Keywords

Cite

@article{arxiv.2505.19931,
  title  = {Accelerating Flow-Matching-Based Text-to-Speech via Empirically Pruned Step Sampling},
  author = {Qixi Zheng and Yushen Chen and Zhikang Niu and Ziyang Ma and Xiaofei Wang and Kai Yu and Xie Chen},
  journal= {arXiv preprint arXiv:2505.19931},
  year   = {2025}
}
R2 v1 2026-07-01T02:39:28.116Z