<link rel="stylesheet" href="styles.f3b1fba60ec7970c.css">

FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech

Loading...
Thumbnail Image

Date

Journal Title

Journal ISSN

Volume Title

Publisher

SLT / Hugging Face

DOI

10.48550/arXiv.2205.12446

Abstract

Speech version of FLoRes with ~2,009 n-way parallel sentences per language read by native speakers, for ASR, ST and language ID evaluation.

Description

HF metadata license: cc-by-4.0; redistribution permitted with attribution. Rights classification: FULL_TEXT_OK. Licence verified from: https://huggingface.co/api/datasets/google/fleurs (checked 2026-11-18). Source: Meta AI (FAIR) African relevance: Includes multiple African languages; key ASR evaluation benchmark. HF metadata license: cc-by-4.0; redistribution permitted with attribution.

Citation

Endorsement

Review

Supplemented By

Referenced By