Multi-lingual audio recognition service dataset

Overtone's APIs allow customers to use our state of the art machine learning and large language models to assess online content. Our models find relevance and significance per business use case, by looking for the kinds of signals that a trained human would too. With our system, we can assess audio, video, speech, dialogue and more and output STT text with annotations. These annotations include entity, sentiment, diarization, tonality, level of opinion and news-worthiness. More services are being adding continuously. We can provide this product for Arabic, English, and other languages.

搜集汇总

背景与挑战

背景概述

该数据集支持多语言音频识别服务，利用先进的机器学习与大型语言模型分析音频、视频等在线内容，输出带实体、情感等标注的STT文本。目前涵盖阿拉伯语、英语等多种语言，且服务功能持续扩展中。