Datasets:

voidful
/

agent-sft-stitch-zh

dataset_info:
  features:
    - name: id
      dtype: large_string
    - name: source
      dtype: large_string
    - name: user
      dtype: large_string
    - name: msg
      dtype: large_string
    - name: input
      dtype: large_string
    - name: total_score
      dtype: int64
    - name: speech_first_score
      dtype: int64
    - name: tool_waiting_safety_score
      dtype: int64
    - name: temporal_causality_score
      dtype: int64
    - name: incremental_update_score
      dtype: int64
    - name: stitch_markup_score
      dtype: int64
    - name: silence_gap_score
      dtype: int64
    - name: grounding_score
      dtype: int64
    - name: naturalness_score
      dtype: int64
    - name: evaluation_json
      dtype: large_string
  splits:
    - name: train
      num_bytes: 216966931
      num_examples: 30502
  download_size: 70307834
  dataset_size: 216966931
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*