---
dataset_info:
  features:
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: images
    list: image
  - name: data_source
    dtype: string
  - name: reward_model
    struct:
    - name: ground_truth
      dtype: string
  - name: extra_info
    struct:
    - name: answer
      dtype: string
    - name: figure_id
      dtype: string
    - name: figure_path
      dtype: string
    - name: index
      dtype: string
    - name: question
      dtype: string
  - name: ability
    dtype: string
  - name: env_name
    dtype: string
  splits:
  - name: train
    num_bytes: 38482239
    num_examples: 304
  download_size: 32386810
  dataset_size: 38482239
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---