---
title: README
emoji: 👀
colorFrom: indigo
colorTo: yellow
sdk: static
pinned: false
---

<div class="grid lg:grid-cols-3 gap-x-4 gap-y-7">
	<p class="lg:col-span-3">
XTREME-S is a benchmark for evaluating universal cross-lingual speech representations in many languages. XTREME-S consists of:
  </p>
	<a
		href="https://arxiv.org/abs/2203.10752"
		class="block overflow-hidden"
	>
		<div class="flex items-center h-40 rounded-lg px-4 mb-2" style="background-color: #FAFBFC;">
			<pre
				class="break-words leading-1 whitespace-pre-line text-xl text-gray-800">
	📚 Paper
		</pre>
		</div>
	</a>
	<a href="https://huggingface.co/datasets/xtreme-s/datasets" class="block overflow-hidden group">
		<div class="flex items-center h-40 rounded-lg px-4 mb-2" style="background-color: #FAFBFC;">
			<pre
				class="break-words leading-1 whitespace-pre-line text-xl text-gray-800">
	🤗 Datasets
		</pre>
		</div>
	</a>
	<a
		href="https://huggingface.co/spaces/xtreme-s/leaderboard"
		class="block overflow-hidden group"
	>
		<div class="flex items-center h-40 rounded-lg px-4 mb-2" style="background-color: #FAFBFC;">
			<pre
				class="break-words leading-1 whitespace-pre-line text-xl text-gray-800">
	🏆 Leaderboard
		</pre>
		</div>
	</a>
    <p class="lg:col-span-3">
        XTREME-S covers four task families: speech recognition, classification, speech-to-text translation and retrieval. Covering 102 languages from 10+ language families, 3 different domains and 4 task families, XTREME-S aims to simplify multilingual speech representation evaluation, as well as catalyze research in "universal" speech representation learning.

        XTREME-S was proposed in the paper *XTREME-S: Evaluating Cross-lingual Speech Representations* by Conneua et. al. in 2022.
	</p>
</div>