Is this your dataset? Claim it with the Hugging Face account that owns it.

speechcolab/
gigaspeech

Name: gigaspeech
Creator: speechcolab
License: https://spdx.org/licenses/Apache-2.0.html

automatic-speech-recognition

Views

No views yet

	Dataset Card for Gigaspeech




	
	Dataset Description

GigaSpeech is an evolving, multi-domain English speech recognition corpus with 10,000 hours of high quality labeled audio suitable for supervised training. The transcribed audio data is collected from audiobooks, podcasts and YouTube, covering both read and spontaneous speaking styles, and a variety of topics, such as arts, science, sports, etc.

Example Usage

The training split has several configurations of… See the full description on the dataset page: https://huggingface.co/datasets/speechcolab/gigaspeech.