README.md

---
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-*
---