ggfox00000/stt-summre-fr-test

SUMM-RE法语测试集是一个用于法语会议风格对话自动语音识别（ASR）和语音活动检测（VAD）的测试数据集。它包含124个单独的音频轨道，来自37个不同的会议，每个会议有3-4个麦克风录音。音频格式为WAV 16 kHz mono PCM_16，嵌入在parquet文件中。数据集包含手动转录和单词级别的强制对齐注释，语言为法语，内容为会议对话。总时长约为41小时，许可证为CC-BY-SA-4.0。

The SUMM-RE French test split is a test dataset for French meeting-style conversational automatic speech recognition (ASR) and voice activity detection (VAD). It contains 124 individual audio tracks from 37 distinct meetings, each with 3-4 microphone recordings. The audio is in WAV 16 kHz mono PCM_16 format, embedded in parquet files. The dataset includes manual transcriptions and word-level forced alignments. The language is French, and the content consists of meeting conversations. The total duration is approximately 41 hours, and the license is CC-BY-SA-4.0.