metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: labels_params
struct:
- name: entity1
dtype: string
- name: entity2
dtype: string
- name: bboxes
struct:
- name: entity1
sequence: float64
- name: entity2
sequence: float64
splits:
- name: train
num_bytes: 37899
num_examples: 300
download_size: 7732
dataset_size: 37899
configs:
- config_name: default
data_files:
- split: train
path: data/train-*