metadata
dataset_info:
features:
- name: transcription
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
splits:
- name: train
num_bytes: 1947609729.4653895
num_examples: 6169
- name: test
num_bytes: 864278563.4406104
num_examples: 2645
download_size: 2705520657
dataset_size: 2811888292.906
license: apache-2.0
Dataset audio info
- 16000 Hz 16 bit
- wav
- mono
- Russian speech
Dataset instance structure
{'audio': {'path': '/path/to/wav.wav', 'array': array([wav numpy array]), dtype=float32), 'sampling_rate': 16000}, 'transcription': 'транскрипция'}
Citation
@Misc{Voxforge.org, author = {Voxforge.org}, title = {Free Speech... Recognition (Linux, Windows and Mac) - voxforge.org}, howpublished = {\url{[http://www.voxforge.org/]}}, note = {accessed 01/21/2023} }