Regotron: Regularizing The Tacotron2 Architecture Via Monotonic Alignment Loss

Abstract

Recent deep learning Text-to-Speech (TTS) systems have achieved impressive performance by generating speech close to human parity. However, they suffer from training stability issues as well as incorrect alignment of the intermediate acoustic representation with the input text sequence. In this work, we introduce Regotron, a regularized version of Tacotron2 which aims to alleviate the training issues and at the same time produce monotonic alignments. Our method augments the vanilla Tacotron2 objective function with an additional term, which penalizes non-monotonic alignments in the location-sensitive attention mechanism. By properly adjusting this regularization term we show that the loss curves become smoother, and at the same time Regotron consistently produces monotonic alignments in unseen examples even at an early stage (13% of the total number of epochs) of its training process, whereas the fully converged Tacotron2 fails to do so. Moreover, our proposed regularization method has

Stats

• citations4
• S2 citations—
• github stars0
• HF likes0
• heat score5.24
• arxiv keygeorgiou2022regotron

Related papers

• Non-attentive Tacotron: Robust And Controllable Neural TTS Synthesis Including Unsupervised Duration Modeling (2020)0.00
• Parallel Tacotron: Non-autoregressive And Controllable TTS (2020)12.54
• Improving Robustness Of Llm-based Speech Synthesis By Learning Monotonic Alignment (2024)0.00
• Efficienttts: An Efficient And High-quality Text-to-speech Architecture (2020)0.00
• Teacher-student Training For Robust Tacotron-based TTS (2019)10.74
• Robust Sequence-to-sequence Acoustic Modeling With Stepwise Monotonic Attention For Neural TTS (2019)11.49
• Tacotron: Towards End-to-end Speech Synthesis (2017)0.00
• Parallel Tacotron 2: A Non-autoregressive Neural TTS Model With Differentiable Duration Modeling (2021)12.68