metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: dev_clean
path: data/dev_clean-*
- split: dev_other
path: data/dev_other-*
- split: test_clean
path: data/test_clean-*
- split: test_other
path: data/test_other-*
dataset_info:
features:
- name: utt_id
dtype: string
- name: hyps
sequence: string
- name: att_score
sequence: float64
- name: ctc_score
sequence: float64
- name: score
sequence: float64
- name: ref
dtype: string
- name: lm_score
sequence: float64
splits:
- name: train
num_bytes: 3073751225
num_examples: 281231
- name: dev_clean
num_bytes: 19839669
num_examples: 2703
- name: dev_other
num_bytes: 18981732
num_examples: 2864
- name: test_clean
num_bytes: 19336959
num_examples: 2620
- name: test_other
num_bytes: 19464386
num_examples: 2939
download_size: 879395852
dataset_size: 3151373971