F5R-TTS：基于组相对策略优化的流匹配文本到语音改进
F5R-TTS: Improving Flow-Matching based Text-to-Speech with Group Relative Policy Optimization
AI总结：
F5R-TTS将GRPO强化学习融入流匹配TTS，通过概率重构实现零样本语音克隆，显著降低WER 29.5%并提升SIM 4.6%。
摘要：
我们提出了F5R-TTS，一种新颖的文本到语音（TTS）系统，它将组相对策略优化（GRPO）集成到基于流匹配的架构中。通过将流匹配TTS的确定性输出重构为概率高斯分布，我们的方法实现了强化学习算法的无缝集成。在预训练阶段，我们使用开源数据集训练了一个基于F5-TTS衍生的概率重构流匹配模型。在随后的强化学习（RL）阶段，我们采用GRPO驱动的增强阶段，利用双重奖励指标：通过自动语音识别计算的词错误率（WER）和通过验证模型评估的说话人相似度（SIM）。零样本语音克隆的实验结果表明，与传统基于流匹配的TTS系统相比，F5R-TTS在语音可懂度（WER相对降低29.5%）和说话人相似度（SIM分数相对提高4.6%）方面均取得了显著改进。音频样本可在https://frontierlabs.github.io/F5R获取。
英文摘要：
We present F5R-TTS, a novel text-to-speech (TTS) system that integrates Group Relative Policy Optimization (GRPO) into a flow-matching based architecture. By reformulating the deterministic outputs of flow-matching TTS into probabilistic Gaussian distributions, our approach enables seamless integration of reinforcement learning algorithms. During pretraining, we train a probabilistically reformulated flow-matching based model which is derived from F5-TTS with an open-source dataset. In the subsequent reinforcement learning (RL) phase, we employ a GRPO-driven enhancement stage that leverages dual reward metrics: word error rate (WER) computed via automatic speech recognition and speaker similarity (SIM) assessed by verification models. Experimental results on zero-shot voice cloning demonstrate that F5R-TTS achieves significant improvements in both speech intelligibility (a 29.5% relative reduction in WER) and speaker similarity (a 4.6% relative increase in SIM score) compared to conventional flow-matching based TTS systems. Audio samples are available at https://frontierlabs.github.io/F5R.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
