metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: chunk_id
dtype: string
- name: speaker_gender
dtype: string
- name: noise_level
dtype: string
- name: contains_music
dtype: bool
- name: contains_overlap
dtype: bool
- name: speakers_estimate
dtype: int32
- name: confidence
dtype: float32
- name: is_uzbek_only
dtype: bool
- name: file_uri
dtype: string
splits:
- name: train
num_bytes: 7908980720.96
num_examples: 3508
download_size: 6421562927
dataset_size: 7908980720.96
configs:
- config_name: default
data_files:
- split: train
path: data/train-*