metadata
dataset_info:
features:
- name: filename
dtype:
audio:
sampling_rate: 24000
- name: duration
dtype: float64
- name: transcript
dtype: string
- name: transcript_mms
dtype: string
- name: levenshtein
dtype: float64
- name: client_id
dtype: int64
- name: filesize
dtype: int64
- name: num_words
dtype: int64
splits:
- name: train
num_bytes: 14422104838.376398
num_examples: 73627
- name: test
num_bytes: 339533974.9070909
num_examples: 1311
- name: dev
num_bytes: 305177744.8748596
num_examples: 1153
download_size: 13313704410
dataset_size: 15066816558.158348
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: dev
path: data/dev-*