RobotsMali/jeli-asr

Jeli-ASR音频数据集是一个多语种数据集，转换为优化的Arrow格式，确保快速访问和与现代数据工作流的兼容性。它包含Bambara语言的半专家转录和法语翻译的音频样本。每个数据集子集都按配置（jeli-asr-rmai、bam-asr-oza和jeli-asr）组织，并进一步分为训练集和测试集。该数据集旨在用于自动语音识别（ASR）、文本到语音合成（TTS）和翻译任务。数据在马里记录，由griots转录并翻译成法语。

The Jeli-ASR Audio Dataset is a multilingual dataset converted into the optimized Arrow format, ensuring fast access and compatibility with modern data workflows. It contains audio samples in Bambara with semi-expert transcriptions and French translations. Each subset of the dataset is organized by configuration (`jeli-asr-rmai`, `bam-asr-oza`, and `jeli-asr`) and further split into training and testing sets. The dataset is designed for tasks like automatic speech recognition (ASR), text-to-speech synthesis (TTS), and translation. Data was recorded in Mali with griots, then transcribed and translated into French.