amorin-gladia/Accueil_UBS_long_form

该数据集名为Accueil UBS long form，收集了40个来自[Accueil_UBS](https://www.ortolang.fr/market/corpora/sldr000890/v1)数据集的电话对话，旨在促进在真实场景（特别是呼叫中心）中法语自动语音识别系统的评估。数据集包含录音文件和对话的文本转录，所有对话均为法语。数据集结构包括文件路径、唯一标识符、对话轮次的JSON格式列表和完整文本转录。对话轮次以JSON格式表示，包含开始时间、结束时间、说话者ID、说话者类型和文本转录。数据集由南布列塔尼大学在真实条件下录制，涵盖来电者和接待人员之间的对话。

This dataset, named Accueil UBS long form, gathers 40 telephone conversations extracted from the [Accueil_UBS](https://www.ortolang.fr/market/corpora/sldr000890/v1) dataset. The objective is to facilitate the evaluation of automatic speech recognition systems in real-world situations, specifically in call centers and in French. The dataset includes recorded audio files and orthographic transcriptions of the dialogues, all in French. The structure of the dataset includes file paths, unique identifiers, a JSON-formatted list of utterances, and complete text transcriptions. Each utterance is represented in JSON format, including start time, end time, speaker ID, speaker type, and text transcription. The corpus was recorded in real conditions at the University of South Brittany and includes dialogues between callers and reception staff.