SautiDB-Naija
收藏arXiv2021-12-12 更新2024-06-21 收录
下载链接：
https://doi.org/10.5281/zenodo.4561842
下载链接
链接失效反馈官方服务：
资源简介：
SautiDB-Naija是由AI Saturdays Lagos创建的一个新颖的非母语（L2）尼日利亚英语语音数据集，包含来自五种尼日利亚语言（Yorùbá, Ìgbò, È. dó, Efik-Ibibio, 和Igala）的919条录音。数据集通过创建文本提示和使用Angular和Firebase开发的网络应用进行收集，经过去噪和标注性别等后处理。该数据集旨在支持机器学习模型在口音转换或翻译及分类任务中的应用，以改善在线学习体验。
SautiDB-Naija is a novel non-native (L2) Nigerian English speech dataset created by AI Saturdays Lagos. It comprises 919 audio recordings from five Nigerian languages: Yorùbá, Ìgbò, Èdó, Efik-Ibibio, and Igala. The dataset was collected via text prompt creation and a web application developed with Angular and Firebase, and underwent post-processing procedures including denoising and gender annotation. This dataset aims to support machine learning models in accent conversion, translation and classification tasks to improve online learning experiences.
提供机构：
AI Saturdays Lagos, 尼日利亚
创建时间：
2021-12-12
