{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/bidirectional-variational-inference-for-non","title":"Bidirectional Variational Inference for Non-Autoregressive Text-to-Speech","arxiv_id":null,"date":"2021-01-01","proceeding":"ICLR 2021 1","authors":["Yoonhyung Lee","Joongbo Shin","Kyomin Jung"],"abstract":"Although early text-to-speech (TTS) models such as Tacotron 2 have succeeded in generating human-like speech, their autoregressive (AR) architectures have a limitation that they require a lot of time to generate a mel-spectrogram consisting of hundreds of steps. In this paper, we propose a novel non-autoregressive TTS model called BVAE-TTS, which eliminates the architectural limitation and generates a mel-spectrogram in parallel. BVAE-TTS adopts bidirectional-inference variational autoencoder (BVAE) that learns hierarchical latent representations using both bottom-up and top-down paths to increase its expressiveness. To apply BVAE to TTS, we design our model to utilize text information via an attention mechanism. By using attention maps that BVAE-TTS generates, we train a duration predictor so that the model uses the predicted length of each phoneme at inference. In experiments conducted on LJSpeech dataset, we show that our model generates a mel-spectrogram 27 times faster than Tacotron 2 with similar speech quality. Furthermore, our BVAE-TTS outperforms Glow-TTS, which is one of the state-of-the-art non-autoregressive TTS models, in terms of both speech quality and inference speed while having 58% fewer parameters.","url_abs":"https://openreview.net/forum?id=o3iritJHLfO","url_pdf":"https://openreview.net/pdf?id=o3iritJHLfO","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"bidirectional-variational-inference-for-non","repo_url":"https://github.com/LEEYOONHYUNG/BVAE-TTS","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"text-to-speech","task_name":"Text to Speech"},{"task_slug":"variational-inference","task_name":"Variational Inference"},{"task_slug":"text-to-speech-1","task_name":"text-to-speech"}],"methods":[{"method_slug":"activation-normalization","method_name":"Activation Normalization"},{"method_slug":"affine-coupling","method_name":"Affine Coupling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bigru","method_name":"BiGRU"},{"method_slug":"bilstm","method_name":"BiLSTM"},{"method_slug":"cbhg","method_name":"CBHG"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dilated-causal-convolution","method_name":"Dilated Causal Convolution"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"glow","method_name":"GLOW"},{"method_slug":"gru","method_name":"GRU"},{"method_slug":"glow-tts","method_name":"Glow-TTS"},{"method_slug":"griffin-lim-algorithm","method_name":"Griffin-Lim Algorithm"},{"method_slug":"highway-layer","method_name":"Highway Layer"},{"method_slug":"highway-network","method_name":"Highway Network"},{"method_slug":"invertible-1x1-convolution","method_name":"Invertible 1x1 Convolution"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"location-sensitive-attention","method_name":"Location Sensitive Attention"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"mixture-of-logistic-distributions","method_name":"Mixture of Logistic Distributions"},{"method_slug":"normalizing-flows","method_name":"Normalizing Flows"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"residual-gru","method_name":"Residual GRU"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tacotron","method_name":"Tacotron"},{"method_slug":"tacotron-2","method_name":"Tacotron 2"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"},{"method_slug":"wavenet","method_name":"WaveNet"},{"method_slug":"zoneout","method_name":"Zoneout"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}