paper-with-me

홈 › Papers

Blank-regularized CTC for Frame Skipping in Neural Transducer

2023-05-19 · Yifan Yang, Xiaoyu Yang, Liyong Guo, Zengwei Yao, Wei Kang, Fangjun Kuang, Long Lin, Xie Chen, Daniel Povey

Neural Transducer and connectionist temporal classification (CTC) are popular end-to-end automatic speech recognition systems. Due to their frame-synchronous design, blank symbols are introduced to address the length mismatch between acoustic frames and output tokens, which might bring redundant computation. Previous studies managed to accelerate the training and inference of neural Transducers by discarding frames based on the blank symbols predicted by a co-trained CTC. However, there is no guarantee that the co-trained CTC can maximize the ratio of blank symbols. This paper proposes two novel regularization methods to explicitly encourage more blanks by constraining the self-loop of non-blank symbols in the CTC. It is interesting to find that the frame reduction ratio of the neural Transducer can approach the theoretical boundary. Experiments on LibriSpeech corpus show that our proposed method accelerates the inference of neural Transducer by 4 times without sacrificing performance. Our work is open-sourced and publicly available https://github.com/k2-fsa/icefall.

📄 PDF Abstract BibTeX arXiv:2305.11558

Code (1)

k2-fsa/icefall 공식 구현 pytorch

Tasks

Automatic Speech Recognitionspeech-recognitionSpeech Recognition

Similar Papers 제목 키워드 기반

Factorized Blank Thresholding for Improved Runtime Efficiency of Neural Transducers

2022-11-02 · Duc Le, Frank Seide, Yuhao Wang, Yang Li 외

We show how factoring the RNN-T's output distribution can significantly reduce the computation cost and power consumption for on-device ASR inference with no loss in accuracy. With the rise in popularity of neural-transd…

Lightweight Transducer Based on Frame-Level Criterion

2024-09-05 · Genshun Wan, Mengzhi Wang, Tingzhi Mao, Hang Chen 외

The transducer model trained based on sequence-level criterion requires a lot of memory due to the generation of the large probability matrix. We proposed a lightweight transducer model based on frame-level criterion, wh…

Decoderimbalanced classificationSpeech Recognition

FSR: Accelerating the Inference Process of Transducer-Based Models by Applying Fast-Skip Regularization

2021-04-07 · Zhengkun Tian, Jiangyan Yi, Ye Bai, JianHua Tao 외

Transducer-based models, such as RNN-Transducer and transformer-transducer, have achieved great success in speech recognition. A typical transducer model decodes the output sequence conditioned on the current acoustic st…

DecoderPositionspeech-recognitionSpeech Recognition

Multi-blank Transducers for Speech Recognition

2022-11-04 · Hainan Xu, Fei Jia, Somshubra Majumdar, Shinji Watanabe 외

This paper proposes a modification to RNN-Transducer (RNN-T) models for automatic speech recognition (ASR). In standard RNN-T, the emission of a blank symbol consumes exactly one input frame; in our proposed method, we i…

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)speech-recognitionSpeech Recognition

CTC Blank Triggered Dynamic Layer-Skipping for Efficient CTC-based Speech Recognition

2024-01-04 · JunFeng Hou, Peiyao Wang, Jincheng Zhang, Meng Yang 외

Deploying end-to-end speech recognition models with limited computing resources remains challenging, despite their impressive performance. Given the gradual increase in model size and the wide range of model applications…

Knowledge Distillationspeech-recognitionSpeech Recognition