metadata
dataset_info:
features:
- name: uuid
dtype: string
- name: speaker
dtype: string
- name: video_id
dtype: string
- name: chunk_file
dtype: string
- name: audio_file
dtype: string
- name: duration
dtype: float64
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: gemini_transcription
dtype: string
- name: tugstugi_transcription
dtype: string
- name: wer
dtype: float64
- name: cer
dtype: float64
- name: ref_words
dtype: int64
- name: hyp_words
dtype: int64
- name: decision
dtype: string
- name: failure_reason
dtype: string
- name: processed_index
dtype: int64
- name: timestamp
dtype: string
splits:
- name: train_0_4
num_bytes: 161165
num_examples: 5
- name: train_5_9
num_bytes: 161165
num_examples: 5
download_size: 19618
dataset_size: 322330
configs:
- config_name: default
data_files:
- split: train_0_4
path: data/train_0_4-*
- split: train_5_9
path: data/train_5_9-*