hf-audio/open-asr-leaderboard-multilingual-datasets

ASR Leaderboard Datasets This repository contains test splits from multiple speech corpora, including FLEURS, Common Voice (MCV), and Multilingual LibriSpeech (MLS). How to Load To load a specific subset, use load_dataset with the corresponding config_name in the format <set>_<lang>. from datasets import load_dataset # Load the FLEURS dataset for Bulgarian fleurs_bg = load_dataset("nithinraok/asr-leaderboard-datasets", "fleurs_bg") print(fleurs_bg) # Load the… See the full description on the dataset page: https://huggingface.co/datasets/hf-audio/open-asr-leaderboard-multilingual-datasets.