BriefGPT.xyz
Jun, 2023
带标点的端到端流式自动语音识别模型的改进训练
Improved Training for End-to-End Streaming Automatic Speech Recognition Model with Punctuation
HTML
PDF
Hanbyul Kim, Seunghyun Seo, Lukas Lee, Seolki Baek
TL;DR
本文提出了一种基于Transformer编码器和CTC loss的方法，实现对输入语音的标点文本进行预测，并通过对文本分块和话语的CTC损失组合，提高了标点预测的准确性和单词错误率。
Abstract
punctuated text prediction
is crucial for
automatic speech recognition
as it enhances readability and impacts downstream
natural language process
→