metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 32000
- name: others
dtype: audio
- name: metadata
dtype: string
- name: input_ids
sequence:
sequence: int32
- name: audio_input_features
sequence:
sequence:
sequence: float32
- name: target_length
dtype: int64
splits:
- name: train
num_bytes: 1101103347
num_examples: 250
- name: eval
num_bytes: 446382158
num_examples: 100
- name: other
num_bytes: 1243634924
num_examples: 280
download_size: 2774397064
dataset_size: 2791120429
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*
- split: other
path: data/other-*