asr-africa/ASRAfricaDataEfficiencyBenchmark

Domain:

natural language processing

Record type:

dataset

Creator:

asr

Host:

This dataset is part of the ASR Africa Data Efficiency Benchmark, designed to evaluate the performance of automatic speech recognition (ASR) models in low-resource settings. It consists of unique MP3 audio files paired with corresponding text transcriptions. Each audio sample is accompanied by metadata, including recording environment, duration, and speaker demographic information such as age and gender.

Visit

huggingface.co

Tasks

automatic speech recognition speech processing