README.md

---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: image
    dtype:
      image:
        mode: RGB
  - name: base_question
    dtype: string
  - name: base_answer
    dtype: string
  - name: subquestions
    dtype: string
  - name: new_subquestions
    dtype: string
  - name: final_question_base
    dtype: string
  - name: response_base
    dtype: string
  - name: ground_truth_base
    dtype: string
  - name: final_ans_base
    dtype: string
  - name: final_question_subques
    dtype: string
  - name: response_subques
    dtype: string
  - name: ground_truth_subques
    dtype: string
  - name: final_ans_subques
    dtype: string
  - name: final_question_chat
    dtype: string
  - name: response_chat
    dtype: string
  - name: ground_truth_chat
    dtype: string
  - name: final_ans_chat
    dtype: string
  - name: split_name
    dtype: string
  - name: image_name
    dtype: string
  - name: confidence_base
    dtype: string
  - name: confidence_subques
    dtype: string
  - name: confidence_chat
    dtype: string
  - name: all_answers_base
    list: string
  - name: all_answers_subques
    list: string
  - name: all_answers_chat
    list: string
  - name: difficulty_our_base
    dtype: int64
  - name: difficulty_our_subques
    dtype: int64
  - name: difficulty_our_chat
    dtype: int64
  - name: difficulty_rationale_our_base
    dtype: string
  - name: difficulty_rationale_our_subques
    dtype: string
  - name: difficulty_rationale_our_chat
    dtype: string
  - name: difficulty_base
    dtype: int64
  - name: rationale_base
    dtype: string
  splits:
  - name: train
    num_bytes: 61569129
    num_examples: 540
  download_size: 57679172
  dataset_size: 61569129
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---