CrowdSpeech: A Benchmark Dataset for Crowdsourced Audio Transcription

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

Models trained or fine-tuned on toloka/CrowdSpeech