metadata
dataset_info:
features:
- name: filename
dtype: string
- name: filesize
dtype: int64
- name: transcript
dtype: string
- name: transcript_wav2vec
dtype: string
- name: levenshtein
dtype: float64
- name: duration
dtype: float64
- name: num_words
dtype: int64
- name: client_id
dtype: int64
- name: audio
dtype:
audio:
sampling_rate: 24000
splits:
- name: train
num_bytes: 5759255141.616
num_examples: 18719
- name: test
num_bytes: 178116894
num_examples: 814
- name: dev
num_bytes: 179299703
num_examples: 853
download_size: 4218674882
dataset_size: 6116671738.616
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: dev
path: data/dev-*