---
annotations_creators:
- found
language_creators:
- crowdsourced
language:
- en-US
license:
- cc-by-4.0
multilinguality:
- monolingual
pretty_name: CrowdSpeech
size_categories:
- unknown
source_datasets:
- original
task_categories:
- automatic-speech-recognition
- conditional-text-generation
task_ids:
- speech-recognition
- summarization
- other-stuctured-to-text
paperswithcode_id: crowdspeech
---

# CrowdSpeech: A Benchmark Dataset for Crowdsourced Audio Transcription

CrowdSpeech — the first publicly available large-scale dataset of crowdsourced audio transcriptions.