SqueezeWave: Extremely Lightweight Vocoders for On-device Speech Synthesis

Automatic speech synthesis is a challenging task that is becoming increasingly important as edge devices begin to interact with users through speech. Typical text-to-speech pipelines include a vocoder, which translates intermediate audio representations into an audio waveform. Most existing vocoders are difficult to parallelize since each generated sample is conditioned on previous samples. WaveGlow is a flow-based feed-forward alternative to these auto-regressive models (Prenger et al., 2019). However, while WaveGlow can be easily parallelized, the model is too expensive for real-time speech synthesis on the edge. This paper presents SqueezeWave, a family of lightweight vocoders based on WaveGlow that can generate audio of similar quality to WaveGlow with 61x - 214x fewer MACs. Code, trained models, and generated audio are publicly available at this https URL.

[1]  Luciano Lavagno,et al.  Synetgy: Algorithm-hardware Co-design for ConvNet Accelerators on Embedded FPGAs , 2018, FPGA.

[2]  Erich Elsen,et al.  Efficient Neural Audio Synthesis , 2018, ICML.

[3]  Sercan Ömer Arik,et al.  Deep Voice 3: 2000-Speaker Neural Text-to-Speech , 2017, ICLR 2018.

[4]  Heiga Zen,et al.  WaveNet: A Generative Model for Raw Audio , 2016, SSW.

[5]  Bo Chen,et al.  MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications , 2017, ArXiv.

[6]  Jan Skoglund,et al.  LPCNET: Improving Neural Speech Synthesis through Linear Prediction , 2018, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[7]  Kurt Keutzer,et al.  SqueezeNext: Hardware-Aware Neural Network Design , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[8]  Yuandong Tian,et al.  FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[9]  Kurt Keutzer,et al.  SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud , 2017, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[10]  Bichen Wu,et al.  Efficient Deep Neural Networks , 2019, ArXiv.

[11]  Navdeep Jaitly,et al.  Natural TTS Synthesis by Conditioning Wavenet on MEL Spectrogram Predictions , 2017, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[12]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[13]  Kurt Keutzer,et al.  Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutions , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[14]  Ryan Prenger,et al.  Waveglow: A Flow-based Generative Network for Speech Synthesis , 2018, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[15]  Kurt Keutzer,et al.  SqueezeSegV2: Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloud , 2018, 2019 International Conference on Robotics and Automation (ICRA).

[16]  Forrest N. Iandola,et al.  SqueezeDet: Unified, Small, Low Power Fully Convolutional Neural Networks for Real-Time Object Detection for Autonomous Driving , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[17]  Forrest N. Iandola,et al.  SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1MB model size , 2016, ArXiv.