metadata
dataset_info:
features:
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: images
list: image
- name: ability
dtype: string
- name: env_name
dtype: string
- name: reward_model
struct:
- name: ground_truth
dtype: string
- name: style
dtype: string
- name: extra_info
struct:
- name: answer
dtype: string
- name: index
dtype: string
- name: question
dtype: string
- name: question_type
dtype: string
- name: plot_id
dtype: string
- name: question_id
dtype: string
- name: split
dtype: string
- name: target_subplots
dtype: string
- name: target_subplot_titles
dtype: string
- name: conditions
dtype: string
- name: qualified_point_ids
dtype: string
- name: qualified_points
dtype: string
- name: reasoning_info
dtype: string
- name: point_coordinates
dtype: string
- name: subplot_boundaries
dtype: string
- name: ground_truth_points
dtype: string
splits:
- name: train
num_bytes: 68326519
num_examples: 3000
download_size: 5494187
dataset_size: 68326519
configs:
- config_name: default
data_files:
- split: train
path: data/train-*