---
dataset_info:
  features:
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: text_instruction
    dtype: string
  - name: data_index
    dtype: string
  - name: relevance_score
    dtype: int32
  - name: quality_score
    dtype: int32
  - name: relevance_reasoning
    dtype: string
  - name: quality_reasoning
    dtype: string
  - name: segments_json
    dtype: string
  - name: origin_model
    dtype: string
  - name: origin_config
    dtype: string
  splits:
  - name: train
    num_bytes: 5355141371.0
    num_examples: 17273
  download_size: 5255098884
  dataset_size: 5355141371.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---