metadata
dataset_info:
features:
- name: image
sequence: image
- name: phy_prompt
dtype: string
- name: non_phy_prompt
dtype: string
- name: phy_summary
dtype: string
- name: non_phy_summary
dtype: string
splits:
- name: train
num_bytes: 24874433
num_examples: 300
download_size: 24736566
dataset_size: 24874433
configs:
- config_name: default
data_files:
- split: train
path: data/train-*