Can I read Fine-Tuning Convergence Model in Bengali Speech Recognition: Zhu Ruiying, Shen Meng on EtoBox?

Fine-Tuning Convergence Model in Bengali Speech Recognition: Zhu Ruiying, Shen Meng by tihisak925 is a document available to read on EtoBox.

What is Fine-Tuning Convergence Model in Bengali Speech Recognition: Zhu Ruiying, Shen Meng about?

This document discusses the fine-tuning of a Bengali speech recognition model utilizing the wave2vec 2.0 architecture, which has shown significant improvements in Word Error Rate (WER) from 0.508 to 0.436. The model was trained on the OOD-Speech dataset and incorporated various optimization strategies, including dropout adjustments and learning rate modifications, to enhance performance and mitigate overfitting. The research aims to advance automatic speech recognition for Bengali, addressing the challenges

Author

tihisak925

Language

EN