metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
splits:
- name: train
num_bytes: 1118863437.65
num_examples: 10226
download_size: 1022411388
dataset_size: 1118863437.65
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Zeroth-STT-Korean Dataset
Description
This is a one of ten equally divided parts of the Zeroth-STT-Korean dataset. Ten equal shards of the original dataset were created and uploaded as individual repos for fine-tuning/training needs.
Citation
Zeroth-Korean Dataset, created by [Lucas Jo(@Atlas Guide Inc.) and Wonkyum Lee(@Gridspace Inc.)], 2023. Available at https://github.com/goodatlas/zeroth under CC-BY-4.0 license.
Junhoee/STT_Korean_Dataset_80000 Dataset, created by [Junhoee], 2024. Available at https://huggingface.co/datasets/Junhoee/STT_Korean_Dataset_80000