ggfox00000/stt-fleurs-test

FLEURS法语测试分割数据集是Google FLEURS项目的一部分，专门用于法语（fr_fr）的自动语音识别（ASR）任务。数据集包含676条语音样本，总时长为1.95小时，音频格式为16 kHz单声道WAV，嵌入在parquet文件中。每条样本包含音频文件、ID、原始转录（带标点）、标准化转录（小写）、字形分割、样本数量和性别信息。数据集的许可证为CC-BY-4.0，来源于Google的FLEURS上游项目。

The FLEURS French test split dataset is part of the Google FLEURS project, specifically designed for automatic speech recognition (ASR) tasks in French (fr_fr). The dataset contains 676 speech samples with a total duration of 1.95 hours, in 16 kHz mono WAV format embedded in parquet files. Each sample includes audio files, ID, raw transcription (with punctuation), normalized transcription (lowercase), grapheme segmentation, number of samples, and gender information. The dataset is licensed under CC-BY-4.0 and sourced from the upstream Google FLEURS project.