arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.09041cs.SDcs.AIeess.AS
收藏

DSFlow:双监督与步感知架构用于一步流匹配语音合成

DSFlow: Dual Supervision and Step-Aware Architecture for One-Step Flow Matching Speech Synthesis

Bin Lin, Peng Yang, Chao Yan, Xiaochen Liu, Wei Wang, Boyong Wu, Pengfei Tan, Xuerui Yang

AI总结:

DSFlow通过双监督和步感知架构提升流匹配语音合成的生成稳定性与参数效率。

摘要:

流匹配模型已经使高质量的文本到语音合成成为可能,但其在推理过程中的迭代采样过程带来了显著的计算成本。尽管蒸馏被广泛用于减少推理步骤的数量,但现有方法常常由于端点误差累积而出现过程变异。此外,直接重用连续时间架构进行离散、固定步长生成会引入结构参数效率低下问题。为了解决这些挑战,我们引入了DSFlow,一种用于少步和一步合成的模块化蒸馏框架。DSFlow将生成重新表述为离散预测任务,并明确将学生模型适应到目标推理制度。它通过结合端点匹配与确定性均速对齐的双监督策略来提高训练稳定性,强制在推理步骤中保持一致的生成轨迹。此外,DSFlow通过将连续时间时间步条件替换为轻量级的步感知标记来提高参数效率,使模型容量与显著减少的时间步空间相匹配。在各种基于流的文本到语音架构上进行的广泛实验表明,DSFlow在标准蒸馏方法上表现一致,实现了强大的少步和一步合成质量,同时减少了模型参数和推理成本。

英文摘要:

Flow-matching models have enabled high-quality text-to-speech synthesis, but their iterative sampling process during inference incurs substantial computational cost. Although distillation is widely used to reduce the number of inference steps, existing methods often suffer from process variance due to endpoint error accumulation. Moreover, directly reusing continuous-time architectures for discrete, fixed-step generation introduces structural parameter inefficiencies. To address these challenges, we introduce DSFlow, a modular distillation framework for few-step and one-step synthesis. DSFlow reformulates generation as a discrete prediction task and explicitly adapts the student model to the target inference regime. It improves training stability through a dual supervision strategy that combines endpoint matching with deterministic mean-velocity alignment, enforcing consistent generation trajectories across inference steps. In addition, DSFlow improves parameter efficiency by replacing continuous-time timestep conditioning with lightweight step-aware tokens, aligning model capacity with the significantly reduced timestep space of the discrete task. Extensive experiments across diverse flow-based text-to-speech architectures demonstrate that DSFlow consistently outperforms standard distillation approaches, achieving strong few-step and one-step synthesis quality while reducing model parameters and inference cost.

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。

↑