metadata
dataset_info:
config_name: estonia_0.08_cer_filtered
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: key
dtype: string
- name: country
dtype: string
- name: language
dtype: string
- name: video_id
dtype: string
- name: transcript_id
dtype: string
- name: start_seconds
dtype: float32
- name: end_seconds
dtype: float32
- name: duration_seconds
dtype: float32
- name: asr_transcript
dtype: string
- name: human_transcript
dtype: string
- name: cer
dtype: float32
- name: wer
dtype: float32
- name: original_transcript_start_idx
dtype: int32
- name: original_transcript_end_idx
dtype: int32
splits:
- name: train
num_bytes: 27132107708.376
num_examples: 56888
- name: test
num_bytes: 473696791.606
num_examples: 1006
- name: validation
num_bytes: 361164537
num_examples: 761
download_size: 27789812690
dataset_size: 27966969036.982
configs:
- config_name: estonia_0.08_cer_filtered
data_files:
- split: train
path: estonia_0.08_cer_filtered/train-*
- split: test
path: estonia_0.08_cer_filtered/test-*
- split: validation
path: estonia_0.08_cer_filtered/validation-*